1 /* Tags file maker to go with GNU Emacs -*- coding: latin-1 -*-
2 Copyright (C) 1984, 87, 88, 89, 93, 94, 95, 98, 99, 2000, 2001
3 Free Software Foundation, Inc. and Ken Arnold
5 This file is not considered part of GNU Emacs.
7 This program is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 2 of the License, or
10 (at your option) any later version.
12 This program is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
17 You should have received a copy of the GNU General Public License
18 along with this program; if not, write to the Free Software Foundation,
19 Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
23 * Ctags originally by Ken Arnold.
24 * Fortran added by Jim Kleckner.
25 * Ed Pelegri-Llopart added C typedefs.
26 * Gnu Emacs TAGS format and modifications by RMS?
27 * 1989 Sam Kendall added C++.
28 * 1993 Francesco Potortì reorganised C and C++ based on work by Joe Wells.
29 * 1994 Regexp tags by Tom Tromey.
30 * 2001 Nested classes by Francesco Potortì based on work by Mykola Dzyuba.
32 * Francesco Potortì <pot@gnu.org> has maintained it since 1993.
35 char pot_etags_version
[] = "@(#) pot revision number is 14.21";
45 # define NDEBUG /* disable assert */
48 #if defined(__STDC__) && (__STDC__ || defined(__SUNPRO_C))
49 # define P_(proto) proto
56 /* On some systems, Emacs defines static as nothing for the sake
57 of unexec. We don't want that here since we don't use unexec. */
59 # define ETAGS_REGEXPS /* use the regexp features */
60 # define LONG_OPTIONS /* accept long options */
63 # define static /* remove static for old compilers' sake */
65 #endif /* !HAVE_CONFIG_H */
68 # define _GNU_SOURCE 1 /* enables some compiler checks on GNU */
71 /* WIN32_NATIVE is for Xemacs.
72 MSDOS, WINDOWSNT, DOS_NT are for Emacs. */
77 #endif /* WIN32_NATIVE */
83 # include <sys/param.h>
85 # ifndef HAVE_CONFIG_H
87 # include <sys/config.h>
99 # define MAXPATHLEN _MAX_PATH
105 # endif /* undef HAVE_GETCWD */
106 #else /* !WINDOWSNT */
111 extern char *getenv ();
113 #endif /* !WINDOWSNT */
118 # if defined (HAVE_GETCWD) && !defined (WINDOWSNT)
119 extern char *getcwd (char *buf
, size_t size
);
121 #endif /* HAVE_UNISTD_H */
129 #include <sys/types.h>
130 #include <sys/stat.h>
134 # undef assert /* some systems have a buggy assert.h */
135 # define assert(x) ((void) 0)
138 #if !defined (S_ISREG) && defined (S_IFREG)
139 # define S_ISREG(m) (((m) & S_IFMT) == S_IFREG)
145 # define getopt_long(argc,argv,optstr,lopts,lind) getopt (argc, argv, optstr)
147 extern int optind
, opterr
;
148 #endif /* LONG_OPTIONS */
152 #endif /* ETAGS_REGEXPS */
154 /* Define CTAGS to make the program "ctags" compatible with the usual one.
155 Leave it undefined to make the program "etags", which makes emacs-style
156 tag tables and tags typedefs, #defines and struct/union/enum by default. */
164 /* Exit codes for success and failure. */
173 #define streq(s,t) (assert((s)!=NULL || (t)!=NULL), !strcmp (s, t))
174 #define strneq(s,t,n) (assert((s)!=NULL || (t)!=NULL), !strncmp (s, t, n))
176 #define CHARS 256 /* 2^sizeof(char) */
177 #define CHAR(x) ((unsigned int)(x) & (CHARS - 1))
178 #define iswhite(c) (_wht[CHAR(c)]) /* c is white */
179 #define notinname(c) (_nin[CHAR(c)]) /* c is not in a name */
180 #define begtoken(c) (_btk[CHAR(c)]) /* c can start token */
181 #define intoken(c) (_itk[CHAR(c)]) /* c can be in token */
182 #define endtoken(c) (_etk[CHAR(c)]) /* c ends tokens */
184 #define ISALNUM(c) isalnum (CHAR(c))
185 #define ISALPHA(c) isalpha (CHAR(c))
186 #define ISDIGIT(c) isdigit (CHAR(c))
187 #define ISLOWER(c) islower (CHAR(c))
189 #define lowcase(c) tolower (CHAR(c))
190 #define upcase(c) toupper (CHAR(c))
194 * xnew, xrnew -- allocate, reallocate storage
196 * SYNOPSIS: Type *xnew (int n, Type);
197 * void xrnew (OldPointer, int n, Type);
200 # include "chkmalloc.h"
201 # define xnew(n,Type) ((Type *) trace_malloc (__FILE__, __LINE__, \
202 (n) * sizeof (Type)))
203 # define xrnew(op,n,Type) ((op) = (Type *) trace_realloc (__FILE__, __LINE__, \
204 (char *) (op), (n) * sizeof (Type)))
206 # define xnew(n,Type) ((Type *) xmalloc ((n) * sizeof (Type)))
207 # define xrnew(op,n,Type) ((op) = (Type *) xrealloc ( \
208 (char *) (op), (n) * sizeof (Type)))
213 typedef void Lang_function
P_((FILE *));
218 char *command
; /* Takes one arg and decompresses to stdout */
224 Lang_function
*function
;
230 typedef struct node_st
231 { /* sorting structure */
232 char *name
; /* function or type name */
233 char *file
; /* file name */
234 bool is_func
; /* use pattern or line no */
235 bool been_warned
; /* set if noticed dup */
236 int lno
; /* line number tag is on */
237 long cno
; /* character number line starts on */
238 char *pat
; /* search pattern */
239 struct node_st
*left
, *right
; /* left and right sons */
243 * A `linebuffer' is a structure which holds a line of text.
244 * `readline_internal' reads a line from a stream into a linebuffer
245 * and works regardless of the length of the line.
246 * SIZE is the size of BUFFER, LEN is the length of the string in
247 * BUFFER after readline reads it.
256 /* Many compilers barf on this:
257 Lang_function Ada_funcs;
258 so let's write it this way */
259 static void Ada_funcs
P_((FILE *));
260 static void Asm_labels
P_((FILE *));
261 static void C_entries
P_((int c_ext
, FILE *));
262 static void default_C_entries
P_((FILE *));
263 static void plain_C_entries
P_((FILE *));
264 static void Cjava_entries
P_((FILE *));
265 static void Cobol_paragraphs
P_((FILE *));
266 static void Cplusplus_entries
P_((FILE *));
267 static void Cstar_entries
P_((FILE *));
268 static void Erlang_functions
P_((FILE *));
269 static void Fortran_functions
P_((FILE *));
270 static void Yacc_entries
P_((FILE *));
271 static void Lisp_functions
P_((FILE *));
272 static void Makefile_targets
P_((FILE *));
273 static void Pascal_functions
P_((FILE *));
274 static void Perl_functions
P_((FILE *));
275 static void Postscript_functions
P_((FILE *));
276 static void Prolog_functions
P_((FILE *));
277 static void Python_functions
P_((FILE *));
278 static void Scheme_functions
P_((FILE *));
279 static void TeX_commands
P_((FILE *));
280 static void Texinfo_nodes
P_((FILE *));
281 static void just_read_file
P_((FILE *));
283 static void print_language_names
P_((void));
284 static void print_version
P_((void));
285 static void print_help
P_((void));
286 int main
P_((int, char **));
287 static int number_len
P_((long));
289 static compressor
*get_compressor_from_suffix
P_((char *, char **));
290 static language
*get_language_from_langname
P_((char *));
291 static language
*get_language_from_interpreter
P_((char *));
292 static language
*get_language_from_filename
P_((char *));
293 static int total_size_of_entries
P_((node
*));
294 static long readline
P_((linebuffer
*, FILE *));
295 static long readline_internal
P_((linebuffer
*, FILE *));
296 static void get_tag
P_((char *));
299 static void analyse_regex
P_((char *, bool));
300 static void add_regex
P_((char *, bool, language
*));
301 static void free_patterns
P_((void));
302 #endif /* ETAGS_REGEXPS */
303 static void error
P_((const char *, const char *));
304 static void suggest_asking_for_help
P_((void));
305 void fatal
P_((char *, char *));
306 static void pfatal
P_((char *));
307 static void add_node
P_((node
*, node
**));
309 static void init
P_((void));
310 static void initbuffer
P_((linebuffer
*));
311 static void find_entries
P_((char *, FILE *));
312 static void free_tree
P_((node
*));
313 static void pfnote
P_((char *, bool, char *, int, int, long));
314 static void new_pfnote
P_((char *, int, bool, char *, int, int, long));
315 static void process_file
P_((char *));
316 static void put_entries
P_((node
*));
317 static void takeprec
P_((void));
319 static char *concat
P_((char *, char *, char *));
320 static char *skip_spaces
P_((char *));
321 static char *skip_non_spaces
P_((char *));
322 static char *savenstr
P_((char *, int));
323 static char *savestr
P_((char *));
324 static char *etags_strchr
P_((const char *, int));
325 static char *etags_strrchr
P_((const char *, int));
326 static char *etags_getcwd
P_((void));
327 static char *relative_filename
P_((char *, char *));
328 static char *absolute_filename
P_((char *, char *));
329 static char *absolute_dirname
P_((char *, char *));
330 static bool filename_is_absolute
P_((char *f
));
331 static void canonicalize_filename
P_((char *));
332 static void linebuffer_setlen
P_((linebuffer
*, int));
333 long *xmalloc
P_((unsigned int));
334 long *xrealloc
P_((char *, unsigned int));
337 char searchar
= '/'; /* use /.../ searches */
339 char *tagfile
; /* output file */
340 char *progname
; /* name this program was invoked with */
341 char *cwd
; /* current working directory */
342 char *tagfiledir
; /* directory of tagfile */
343 FILE *tagf
; /* ioptr for tags file */
345 char *curfile
; /* current input file name */
346 language
*curlang
; /* current language */
348 int lineno
; /* line number of current line */
349 long charno
; /* current character number */
350 long linecharno
; /* charno of start of current line */
351 char *dbp
; /* pointer to start of current tag */
353 node
*head
; /* the head of the binary tree of tags */
355 linebuffer lb
; /* the current line */
357 /* boolean "functions" (see init) */
358 bool _wht
[CHARS
], _nin
[CHARS
], _itk
[CHARS
], _btk
[CHARS
], _etk
[CHARS
];
361 *white
= " \f\t\n\r\v",
363 *nonam
= " \f\t\n\r(=,[;",
364 /* token ending chars */
365 *endtk
= " \t\n\r\"'#()[]{}=-+%*/&|^~!<>;,.:?",
366 /* token starting chars */
367 *begtk
= "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$~@",
368 /* valid in-token chars */
369 *midtk
= "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$0123456789";
371 bool append_to_tagfile
; /* -a: append to tags */
372 /* The following four default to TRUE for etags, but to FALSE for ctags. */
373 bool typedefs
; /* -t: create tags for C and Ada typedefs */
374 bool typedefs_or_cplusplus
; /* -T: create tags for C typedefs, level */
375 /* 0 struct/enum/union decls, and C++ */
376 /* member functions. */
377 bool constantypedefs
; /* -d: create tags for C #define, enum */
378 /* constants and variables. */
379 /* -D: opposite of -d. Default under ctags. */
380 bool declarations
; /* --declarations: tag them and extern in C&Co*/
381 bool globals
; /* create tags for global variables */
382 bool members
; /* create tags for C member variables */
383 bool update
; /* -u: update tags */
384 bool vgrind_style
; /* -v: create vgrind style index output */
385 bool no_warnings
; /* -w: suppress warnings */
386 bool cxref_style
; /* -x: create cxref style output */
387 bool cplusplus
; /* .[hc] means C++, not C */
388 bool noindentypedefs
; /* -I: ignore indentation in C */
389 bool packages_only
; /* --packages-only: in Ada, only tag packages*/
392 struct option longopts
[] =
394 { "packages-only", no_argument
, &packages_only
, TRUE
},
395 { "append", no_argument
, NULL
, 'a' },
396 { "backward-search", no_argument
, NULL
, 'B' },
397 { "c++", no_argument
, NULL
, 'C' },
398 { "cxref", no_argument
, NULL
, 'x' },
399 { "defines", no_argument
, NULL
, 'd' },
400 { "declarations", no_argument
, &declarations
, TRUE
},
401 { "no-defines", no_argument
, NULL
, 'D' },
402 { "globals", no_argument
, &globals
, TRUE
},
403 { "no-globals", no_argument
, &globals
, FALSE
},
404 { "help", no_argument
, NULL
, 'h' },
405 { "help", no_argument
, NULL
, 'H' },
406 { "ignore-indentation", no_argument
, NULL
, 'I' },
407 { "include", required_argument
, NULL
, 'i' },
408 { "language", required_argument
, NULL
, 'l' },
409 { "members", no_argument
, &members
, TRUE
},
410 { "no-members", no_argument
, &members
, FALSE
},
411 { "no-warn", no_argument
, NULL
, 'w' },
412 { "output", required_argument
, NULL
, 'o' },
414 { "regex", required_argument
, NULL
, 'r' },
415 { "no-regex", no_argument
, NULL
, 'R' },
416 { "ignore-case-regex", required_argument
, NULL
, 'c' },
417 #endif /* ETAGS_REGEXPS */
418 { "typedefs", no_argument
, NULL
, 't' },
419 { "typedefs-and-c++", no_argument
, NULL
, 'T' },
420 { "update", no_argument
, NULL
, 'u' },
421 { "version", no_argument
, NULL
, 'V' },
422 { "vgrind", no_argument
, NULL
, 'v' },
425 #endif /* LONG_OPTIONS */
428 /* Structure defining a regular expression. Elements are
429 the compiled pattern, and the name string. */
430 typedef struct pattern
432 struct pattern
*p_next
;
435 struct re_pattern_buffer
*pattern
;
436 struct re_registers regs
;
441 /* List of all regexps. */
442 pattern
*p_head
= NULL
;
444 /* How many characters in the character set. (From regex.c.) */
445 #define CHAR_SET_SIZE 256
446 /* Translation table for case-insensitive matching. */
447 char lc_trans
[CHAR_SET_SIZE
];
448 #endif /* ETAGS_REGEXPS */
450 compressor compressors
[] =
452 { "z", "gzip -d -c"},
453 { "Z", "gzip -d -c"},
454 { "gz", "gzip -d -c"},
455 { "GZ", "gzip -d -c"},
456 { "bz2", "bzip2 -d -c" },
464 /* Non-NULL if language fixed. */
465 language
*forced_lang
= NULL
;
468 char *Ada_suffixes
[] =
469 { "ads", "adb", "ada", NULL
};
472 char *Asm_suffixes
[] = { "a", /* Unix assembler */
473 "asm", /* Microcontroller assembly */
474 "def", /* BSO/Tasking definition includes */
475 "inc", /* Microcontroller include files */
476 "ins", /* Microcontroller include files */
477 "s", "sa", /* Unix assembler */
478 "S", /* cpp-processed Unix assembler */
479 "src", /* BSO/Tasking C compiler output */
483 /* Note that .c and .h can be considered C++, if the --c++ flag was
484 given, or if the `class' keyowrd is met inside the file.
485 That is why default_C_entries is called for these. */
486 char *default_C_suffixes
[] =
489 char *Cplusplus_suffixes
[] =
490 { "C", "c++", "cc", "cpp", "cxx", "H", "h++", "hh", "hpp", "hxx",
491 "M", /* Objective C++ */
492 "pdb", /* Postscript with C syntax */
495 char *Cjava_suffixes
[] =
498 char *Cobol_suffixes
[] =
499 { "COB", "cob", NULL
};
501 char *Cstar_suffixes
[] =
502 { "cs", "hs", NULL
};
504 char *Erlang_suffixes
[] =
505 { "erl", "hrl", NULL
};
507 char *Fortran_suffixes
[] =
508 { "F", "f", "f90", "for", NULL
};
510 char *Lisp_suffixes
[] =
511 { "cl", "clisp", "el", "l", "lisp", "LSP", "lsp", "ml", NULL
};
513 char *Makefile_filenames
[] =
514 { "Makefile", "makefile", "GNUMakefile", "Makefile.in", "Makefile.am", NULL
};
516 char *Pascal_suffixes
[] =
517 { "p", "pas", NULL
};
519 char *Perl_suffixes
[] =
520 { "pl", "pm", NULL
};
521 char *Perl_interpreters
[] =
522 { "perl", "@PERL@", NULL
};
524 char *plain_C_suffixes
[] =
525 { "lm", /* Objective lex file */
526 "m", /* Objective C file */
527 "pc", /* Pro*C file */
530 char *Postscript_suffixes
[] =
531 { "ps", "psw", NULL
}; /* .psw is for PSWrap */
533 char *Prolog_suffixes
[] =
536 char *Python_suffixes
[] =
539 /* Can't do the `SCM' or `scm' prefix with a version number. */
540 char *Scheme_suffixes
[] =
541 { "oak", "sch", "scheme", "SCM", "scm", "SM", "sm", "ss", "t", NULL
};
543 char *TeX_suffixes
[] =
544 { "bib", "clo", "cls", "ltx", "sty", "TeX", "tex", NULL
};
546 char *Texinfo_suffixes
[] =
547 { "texi", "texinfo", "txi", NULL
};
549 char *Yacc_suffixes
[] =
550 { "y", "y++", "ym", "yxx", "yy", NULL
}; /* .ym is Objective yacc file */
553 * Table of languages.
555 * It is ok for a given function to be listed under more than one
556 * name. I just didn't.
559 language lang_names
[] =
561 { "ada", Ada_funcs
, NULL
, Ada_suffixes
, NULL
},
562 { "asm", Asm_labels
, NULL
, Asm_suffixes
, NULL
},
563 { "c", default_C_entries
, NULL
, default_C_suffixes
, NULL
},
564 { "c++", Cplusplus_entries
, NULL
, Cplusplus_suffixes
, NULL
},
565 { "c*", Cstar_entries
, NULL
, Cstar_suffixes
, NULL
},
566 { "cobol", Cobol_paragraphs
, NULL
, Cobol_suffixes
, NULL
},
567 { "erlang", Erlang_functions
, NULL
, Erlang_suffixes
, NULL
},
568 { "fortran", Fortran_functions
, NULL
, Fortran_suffixes
, NULL
},
569 { "java", Cjava_entries
, NULL
, Cjava_suffixes
, NULL
},
570 { "lisp", Lisp_functions
, NULL
, Lisp_suffixes
, NULL
},
571 { "makefile", Makefile_targets
, Makefile_filenames
, NULL
, NULL
},
572 { "pascal", Pascal_functions
, NULL
, Pascal_suffixes
, NULL
},
573 { "perl", Perl_functions
, NULL
, Perl_suffixes
, Perl_interpreters
},
574 { "postscript", Postscript_functions
, NULL
, Postscript_suffixes
, NULL
},
575 { "proc", plain_C_entries
, NULL
, plain_C_suffixes
, NULL
},
576 { "prolog", Prolog_functions
, NULL
, Prolog_suffixes
, NULL
},
577 { "python", Python_functions
, NULL
, Python_suffixes
, NULL
},
578 { "scheme", Scheme_functions
, NULL
, Scheme_suffixes
, NULL
},
579 { "tex", TeX_commands
, NULL
, TeX_suffixes
, NULL
},
580 { "texinfo", Texinfo_nodes
, NULL
, Texinfo_suffixes
, NULL
},
581 { "yacc", Yacc_entries
, NULL
, Yacc_suffixes
, NULL
},
582 { "auto", NULL
}, /* default guessing scheme */
583 { "none", just_read_file
}, /* regexp matching only */
584 { NULL
, NULL
} /* end of list */
589 print_language_names ()
594 puts ("\nThese are the currently supported languages, along with the\n\
595 default file names and dot suffixes:");
596 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
598 printf (" %-*s", 10, lang
->name
);
599 if (lang
->filenames
!= NULL
)
600 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
601 printf (" %s", *name
);
602 if (lang
->suffixes
!= NULL
)
603 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
604 printf (" .%s", *ext
);
607 puts ("Where `auto' means use default language for files based on file\n\
608 name suffix, and `none' means only do regexp processing on files.\n\
609 If no language is specified and no matching suffix is found,\n\
610 the first line of the file is read for a sharp-bang (#!) sequence\n\
611 followed by the name of an interpreter. If no such sequence is found,\n\
612 Fortran is tried first; if no tags are found, C is tried next.\n\
613 When parsing any C file, a \"class\" keyword switches to C++.\n\
614 Compressed files are supported using gzip and bzip2.");
618 # define EMACS_NAME "GNU Emacs"
621 # define VERSION "21"
626 printf ("%s (%s %s)\n", (CTAGS
) ? "ctags" : "etags", EMACS_NAME
, VERSION
);
627 puts ("Copyright (C) 1999 Free Software Foundation, Inc. and Ken Arnold");
628 puts ("This program is distributed under the same terms as Emacs");
636 printf ("Usage: %s [options] [[regex-option ...] file-name] ...\n\
638 These are the options accepted by %s.\n", progname
, progname
);
640 puts ("You may use unambiguous abbreviations for the long option names.");
642 puts ("Long option names do not work with this executable, as it is not\n\
643 linked with GNU getopt.");
644 #endif /* LONG_OPTIONS */
645 puts ("A - as file name means read names from stdin (one per line).");
647 printf (" Absolute names are stored in the output file as they are.\n\
648 Relative ones are stored relative to the output file's directory.");
651 puts ("-a, --append\n\
652 Append tag entries to existing tags file.");
654 puts ("--packages-only\n\
655 For Ada files, only generate tags for packages .");
658 puts ("-B, --backward-search\n\
659 Write the search commands for the tag entries using '?', the\n\
660 backward-search command instead of '/', the forward-search command.");
662 /* This option is mostly obsolete, because etags can now automatically
663 detect C++. Retained for backward compatibility and for debugging and
664 experimentation. In principle, we could want to tag as C++ even
665 before any "class" keyword.
667 Treat files whose name suffix defaults to C language as C++ files.");
670 puts ("--declarations\n\
671 In C and derived languages, create tags for function declarations,");
673 puts ("\tand create tags for extern variables if --globals is used.");
676 ("\tand create tags for extern variables unless --no-globals is used.");
679 puts ("-d, --defines\n\
680 Create tag entries for C #define constants and enum constants, too.");
682 puts ("-D, --no-defines\n\
683 Don't create tag entries for C #define constants and enum constants.\n\
684 This makes the tags file smaller.");
688 puts ("-i FILE, --include=FILE\n\
689 Include a note in tag file indicating that, when searching for\n\
690 a tag, one should also consult the tags file FILE after\n\
691 checking the current file.");
692 puts ("-l LANG, --language=LANG\n\
693 Force the following files to be considered as written in the\n\
694 named language up to the next --language=LANG option.");
699 Create tag entries for global variables in some languages.");
701 puts ("--no-globals\n\
702 Do not create tag entries for global variables in some\n\
703 languages. This makes the tags file smaller.");
705 Create tag entries for member variables in C and derived languages.");
708 puts ("-r /REGEXP/, --regex=/REGEXP/ or --regex=@regexfile\n\
709 Make a tag for each line matching pattern REGEXP in the following\n\
710 files. {LANGUAGE}/REGEXP/ uses REGEXP for LANGUAGE files only.\n\
711 regexfile is a file containing one REGEXP per line.\n\
712 REGEXP is anchored (as if preceded by ^).\n\
713 The form /REGEXP/NAME/ creates a named tag.\n\
714 For example Tcl named tags can be created with:\n\
715 --regex=\"/proc[ \\t]+\\([^ \\t]+\\)/\\1/.\"");
716 puts ("-c /REGEXP/, --ignore-case-regex=/REGEXP/ or --ignore-case-regex=@regexfile\n\
717 Like -r, --regex but ignore case when matching expressions.");
718 puts ("-R, --no-regex\n\
719 Don't create tags from regexps for the following files.");
720 #endif /* ETAGS_REGEXPS */
721 puts ("-o FILE, --output=FILE\n\
722 Write the tags to FILE.");
723 puts ("-I, --ignore-indentation\n\
724 Don't rely on indentation quite as much as normal. Currently,\n\
725 this means not to assume that a closing brace in the first\n\
726 column is the final brace of a function or structure\n\
727 definition in C and C++.");
731 puts ("-t, --typedefs\n\
732 Generate tag entries for C and Ada typedefs.");
733 puts ("-T, --typedefs-and-c++\n\
734 Generate tag entries for C typedefs, C struct/enum/union tags,\n\
735 and C++ member functions.");
736 puts ("-u, --update\n\
737 Update the tag entries for the given files, leaving tag\n\
738 entries for other files in place. Currently, this is\n\
739 implemented by deleting the existing entries for the given\n\
740 files and then rewriting the new entries at the end of the\n\
741 tags file. It is often faster to simply rebuild the entire\n\
742 tag file than to use this.");
743 puts ("-v, --vgrind\n\
744 Generates an index of items intended for human consumption,\n\
745 similar to the output of vgrind. The index is sorted, and\n\
746 gives the page number of each item.");
747 puts ("-w, --no-warn\n\
748 Suppress warning messages about entries defined in multiple\n\
750 puts ("-x, --cxref\n\
751 Like --vgrind, but in the style of cxref, rather than vgrind.\n\
752 The output uses line numbers instead of page numbers, but\n\
753 beyond that the differences are cosmetic; try both to see\n\
757 puts ("-V, --version\n\
758 Print the version of the program.\n\
760 Print this help message.");
762 print_language_names ();
765 puts ("Report bugs to bug-gnu-emacs@gnu.org");
779 /* This structure helps us allow mixing of --lang and file names. */
782 enum argument_type arg_type
;
784 language
*lang
; /* language of the regexp */
787 #ifdef VMS /* VMS specific functions */
791 /* This is a BUG! ANY arbitrary limit is a BUG!
792 Won't someone please fix this? */
793 #define MAX_FILE_SPEC_LEN 255
796 char body
[MAX_FILE_SPEC_LEN
+ 1];
800 v1.05 nmm 26-Jun-86 fn_exp - expand specification of list of file names
801 returning in each successive call the next file name matching the input
802 spec. The function expects that each in_spec passed
803 to it will be processed to completion; in particular, up to and
804 including the call following that in which the last matching name
805 is returned, the function ignores the value of in_spec, and will
806 only start processing a new spec with the following call.
807 If an error occurs, on return out_spec contains the value
808 of in_spec when the error occurred.
810 With each successive file name returned in out_spec, the
811 function's return value is one. When there are no more matching
812 names the function returns zero. If on the first call no file
813 matches in_spec, or there is any other error, -1 is returned.
818 #define OUTSIZE MAX_FILE_SPEC_LEN
824 static long context
= 0;
825 static struct dsc$descriptor_s o
;
826 static struct dsc$descriptor_s i
;
827 static bool pass1
= TRUE
;
834 o
.dsc$a_pointer
= (char *) out
;
835 o
.dsc$w_length
= (short)OUTSIZE
;
836 i
.dsc$a_pointer
= in
;
837 i
.dsc$w_length
= (short)strlen(in
);
838 i
.dsc$b_dtype
= DSC$K_DTYPE_T
;
839 i
.dsc$b_class
= DSC$K_CLASS_S
;
840 o
.dsc$b_dtype
= DSC$K_DTYPE_VT
;
841 o
.dsc$b_class
= DSC$K_CLASS_VS
;
843 if ((status
= lib$
find_file(&i
, &o
, &context
, 0, 0)) == RMS$_NORMAL
)
845 out
->body
[out
->curlen
] = EOS
;
848 else if (status
== RMS$_NMF
)
852 strcpy(out
->body
, in
);
855 lib$
find_file_end(&context
);
861 v1.01 nmm 19-Aug-85 gfnames - return in successive calls the
862 name of each file specified by the provided arg expanding wildcards.
865 gfnames (arg
, p_error
)
869 static vspec filename
= {MAX_FILE_SPEC_LEN
, "\0"};
871 switch (fn_exp (&filename
, arg
))
875 return filename
.body
;
881 return filename
.body
;
885 #ifndef OLD /* Newer versions of VMS do provide `system'. */
889 error ("%s", "system() function not implemented under VMS");
893 #define VERSION_DELIM ';'
894 char *massage_name (s
)
900 if (*s
== VERSION_DELIM
)
918 unsigned int nincluded_files
;
919 char **included_files
;
922 int current_arg
, file_count
;
923 linebuffer filename_lb
;
929 _fmode
= O_BINARY
; /* all of files are treated as binary files */
934 included_files
= xnew (argc
, char *);
938 /* Allocate enough no matter what happens. Overkill, but each one
940 argbuffer
= xnew (argc
, argument
);
943 /* Set syntax for regular expression routines. */
944 re_set_syntax (RE_SYNTAX_EMACS
| RE_INTERVALS
);
945 /* Translation table for case-insensitive search. */
946 for (i
= 0; i
< CHAR_SET_SIZE
; i
++)
947 lc_trans
[i
] = lowcase (i
);
948 #endif /* ETAGS_REGEXPS */
951 * If etags, always find typedefs and structure tags. Why not?
952 * Also default to find macro constants, enum constants and
957 typedefs
= typedefs_or_cplusplus
= constantypedefs
= TRUE
;
959 declarations
= FALSE
;
969 optstring
= "-aCdDf:Il:o:r:c:RStTi:BuvxwVhH";
971 optstring
= "-aCdDf:Il:o:StTi:BuvxwVhH";
972 #endif /* ETAGS_REGEXPS */
975 optstring
= optstring
+ 1;
976 #endif /* LONG_OPTIONS */
978 opt
= getopt_long (argc
, argv
, optstring
, longopts
, 0);
985 /* If getopt returns 0, then it has already processed a
986 long-named option. We should do nothing. */
990 /* This means that a file name has been seen. Record it. */
991 argbuffer
[current_arg
].arg_type
= at_filename
;
992 argbuffer
[current_arg
].what
= optarg
;
997 /* Common options. */
998 case 'a': append_to_tagfile
= TRUE
; break;
999 case 'C': cplusplus
= TRUE
; break;
1000 case 'd': constantypedefs
= TRUE
; break;
1001 case 'D': constantypedefs
= FALSE
; break;
1002 case 'f': /* for compatibility with old makefiles */
1006 error ("-o option may only be given once.", (char *)NULL
);
1007 suggest_asking_for_help ();
1012 case 'S': /* for backward compatibility */
1013 noindentypedefs
= TRUE
;
1017 language
*lang
= get_language_from_langname (optarg
);
1020 argbuffer
[current_arg
].lang
= lang
;
1021 argbuffer
[current_arg
].arg_type
= at_language
;
1026 #ifdef ETAGS_REGEXPS
1028 argbuffer
[current_arg
].arg_type
= at_regexp
;
1029 argbuffer
[current_arg
].what
= optarg
;
1033 argbuffer
[current_arg
].arg_type
= at_regexp
;
1034 argbuffer
[current_arg
].what
= NULL
;
1038 argbuffer
[current_arg
].arg_type
= at_icregexp
;
1039 argbuffer
[current_arg
].what
= optarg
;
1042 #endif /* ETAGS_REGEXPS */
1054 typedefs
= typedefs_or_cplusplus
= TRUE
;
1059 included_files
[nincluded_files
++] = optarg
;
1062 /* Ctags options. */
1063 case 'B': searchar
= '?'; break;
1064 case 'u': update
= TRUE
; break;
1065 case 'v': vgrind_style
= TRUE
; /*FALLTHRU*/
1066 case 'x': cxref_style
= TRUE
; break;
1067 case 'w': no_warnings
= TRUE
; break;
1070 suggest_asking_for_help ();
1074 for (; optind
< argc
; ++optind
)
1076 argbuffer
[current_arg
].arg_type
= at_filename
;
1077 argbuffer
[current_arg
].what
= argv
[optind
];
1082 if (nincluded_files
== 0 && file_count
== 0)
1084 error ("no input files specified.", (char *)NULL
);
1085 suggest_asking_for_help ();
1088 if (tagfile
== NULL
)
1089 tagfile
= CTAGS
? "tags" : "TAGS";
1090 cwd
= etags_getcwd (); /* the current working directory */
1091 if (cwd
[strlen (cwd
) - 1] != '/')
1094 cwd
= concat (oldcwd
, "/", "");
1097 if (streq (tagfile
, "-"))
1100 tagfiledir
= absolute_dirname (tagfile
, cwd
);
1102 init (); /* set up boolean "functions" */
1105 initbuffer (&filename_lb
);
1109 if (streq (tagfile
, "-"))
1113 /* Switch redirected `stdout' to binary mode (setting `_fmode'
1114 doesn't take effect until after `stdout' is already open). */
1115 if (!isatty (fileno (stdout
)))
1116 setmode (fileno (stdout
), O_BINARY
);
1120 tagf
= fopen (tagfile
, append_to_tagfile
? "a" : "w");
1126 * Loop through files finding functions.
1128 for (i
= 0; i
< current_arg
; ++i
)
1130 switch (argbuffer
[i
].arg_type
)
1133 forced_lang
= argbuffer
[i
].lang
;
1135 #ifdef ETAGS_REGEXPS
1137 analyse_regex (argbuffer
[i
].what
, FALSE
);
1140 analyse_regex (argbuffer
[i
].what
, TRUE
);
1145 while ((this_file
= gfnames (argbuffer
[i
].what
, &got_err
)) != NULL
)
1149 error ("can't find file %s\n", this_file
);
1154 this_file
= massage_name (this_file
);
1157 this_file
= argbuffer
[i
].what
;
1159 /* Input file named "-" means read file names from stdin
1160 (one per line) and use them. */
1161 if (streq (this_file
, "-"))
1162 while (readline_internal (&filename_lb
, stdin
) > 0)
1163 process_file (filename_lb
.buffer
);
1165 process_file (this_file
);
1173 #ifdef ETAGS_REGEXPS
1175 #endif /* ETAGS_REGEXPS */
1179 while (nincluded_files
-- > 0)
1180 fprintf (tagf
, "\f\n%s,include\n", *included_files
++);
1186 /* If CTAGS, we are here. process_file did not write the tags yet,
1187 because we want them ordered. Let's do it now. */
1199 for (i
= 0; i
< current_arg
; ++i
)
1201 if (argbuffer
[i
].arg_type
!= at_filename
)
1204 "mv %s OTAGS;fgrep -v '\t%s\t' OTAGS >%s;rm OTAGS",
1205 tagfile
, argbuffer
[i
].what
, tagfile
);
1206 if (system (cmd
) != GOOD
)
1207 fatal ("failed to execute shell command", (char *)NULL
);
1209 append_to_tagfile
= TRUE
;
1212 tagf
= fopen (tagfile
, append_to_tagfile
? "a" : "w");
1222 char cmd
[2*BUFSIZ
+10];
1223 sprintf (cmd
, "sort -o %.*s %.*s", BUFSIZ
, tagfile
, BUFSIZ
, tagfile
);
1224 exit (system (cmd
));
1232 * Return a compressor given the file name. If EXTPTR is non-zero,
1233 * return a pointer into FILE where the compressor-specific
1234 * extension begins. If no compressor is found, NULL is returned
1235 * and EXTPTR is not significant.
1236 * Idea by Vladimir Alexiev <vladimir@cs.ualberta.ca> (1998)
1239 get_compressor_from_suffix (file
, extptr
)
1244 char *slash
, *suffix
;
1246 /* This relies on FN to be after canonicalize_filename,
1247 so we don't need to consider backslashes on DOS_NT. */
1248 slash
= etags_strrchr (file
, '/');
1249 suffix
= etags_strrchr (file
, '.');
1250 if (suffix
== NULL
|| suffix
< slash
)
1255 /* Let those poor souls who live with DOS 8+3 file name limits get
1256 some solace by treating foo.cgz as if it were foo.c.gz, etc.
1257 Only the first do loop is run if not MSDOS */
1260 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1261 if (streq (compr
->suffix
, suffix
))
1264 break; /* do it only once: not really a loop */
1267 } while (*suffix
!= '\0');
1274 * Return a language given the name.
1277 get_language_from_langname (name
)
1283 error ("empty language name", (char *)NULL
);
1286 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1287 if (streq (name
, lang
->name
))
1289 error ("unknown language \"%s\"", name
);
1297 * Return a language given the interpreter name.
1300 get_language_from_interpreter (interpreter
)
1306 if (interpreter
== NULL
)
1308 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1309 if (lang
->interpreters
!= NULL
)
1310 for (iname
= lang
->interpreters
; *iname
!= NULL
; iname
++)
1311 if (streq (*iname
, interpreter
))
1320 * Return a language given the file name.
1323 get_language_from_filename (file
)
1327 char **name
, **ext
, *suffix
;
1329 /* Try whole file name first. */
1330 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1331 if (lang
->filenames
!= NULL
)
1332 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
1333 if (streq (*name
, file
))
1336 /* If not found, try suffix after last dot. */
1337 suffix
= etags_strrchr (file
, '.');
1341 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1342 if (lang
->suffixes
!= NULL
)
1343 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
1344 if (streq (*ext
, suffix
))
1352 * This routine is called on each file argument.
1358 struct stat stat_buf
;
1361 char *compressed_name
, *uncompressed_name
;
1362 char *ext
, *real_name
;
1365 canonicalize_filename (file
);
1366 if (streq (file
, tagfile
) && !streq (tagfile
, "-"))
1368 error ("skipping inclusion of %s in self.", file
);
1371 if ((compr
= get_compressor_from_suffix (file
, &ext
)) == NULL
)
1373 compressed_name
= NULL
;
1374 real_name
= uncompressed_name
= savestr (file
);
1378 real_name
= compressed_name
= savestr (file
);
1379 uncompressed_name
= savenstr (file
, ext
- file
);
1382 /* If the canonicalised uncompressed name has already be dealt with,
1383 skip it silently, else add it to the list. */
1385 typedef struct processed_file
1388 struct processed_file
*next
;
1390 static processed_file
*pf_head
= NULL
;
1391 register processed_file
*fnp
;
1393 for (fnp
= pf_head
; fnp
!= NULL
; fnp
= fnp
->next
)
1394 if (streq (uncompressed_name
, fnp
->filename
))
1397 pf_head
= xnew (1, struct processed_file
);
1398 pf_head
->filename
= savestr (uncompressed_name
);
1399 pf_head
->next
= fnp
;
1402 if (stat (real_name
, &stat_buf
) != 0)
1404 /* Reset real_name and try with a different name. */
1406 if (compressed_name
!= NULL
) /* try with the given suffix */
1408 if (stat (uncompressed_name
, &stat_buf
) == 0)
1409 real_name
= uncompressed_name
;
1411 else /* try all possible suffixes */
1413 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1415 compressed_name
= concat (file
, ".", compr
->suffix
);
1416 if (stat (compressed_name
, &stat_buf
) != 0)
1420 char *suf
= compressed_name
+ strlen (file
);
1421 size_t suflen
= strlen (compr
->suffix
) + 1;
1422 for ( ; suf
[1]; suf
++, suflen
--)
1424 memmove (suf
, suf
+ 1, suflen
);
1425 if (stat (compressed_name
, &stat_buf
) == 0)
1427 real_name
= compressed_name
;
1431 if (real_name
!= NULL
)
1434 free (compressed_name
);
1435 compressed_name
= NULL
;
1439 real_name
= compressed_name
;
1444 if (real_name
== NULL
)
1449 } /* try with a different name */
1451 if (!S_ISREG (stat_buf
.st_mode
))
1453 error ("skipping %s: it is not a regular file.", real_name
);
1456 if (real_name
== compressed_name
)
1458 char *cmd
= concat (compr
->command
, " ", real_name
);
1459 inf
= (FILE *) popen (cmd
, "r");
1463 inf
= fopen (real_name
, "r");
1470 find_entries (uncompressed_name
, inf
);
1472 if (real_name
== compressed_name
)
1481 if (filename_is_absolute (uncompressed_name
))
1483 /* file is an absolute file name. Canonicalise it. */
1484 filename
= absolute_filename (uncompressed_name
, cwd
);
1488 /* file is a file name relative to cwd. Make it relative
1489 to the directory of the tags file. */
1490 filename
= relative_filename (uncompressed_name
, tagfiledir
);
1492 fprintf (tagf
, "\f\n%s,%d\n", filename
, total_size_of_entries (head
));
1500 if (compressed_name
) free(compressed_name
);
1501 if (uncompressed_name
) free(uncompressed_name
);
1506 * This routine sets up the boolean pseudo-functions which work
1507 * by setting boolean flags dependent upon the corresponding character.
1508 * Every char which is NOT in that string is not a white char. Therefore,
1509 * all of the array "_wht" is set to FALSE, and then the elements
1510 * subscripted by the chars in "white" are set to TRUE. Thus "_wht"
1511 * of a char is TRUE if it is the string "white", else FALSE.
1519 for (i
= 0; i
< CHARS
; i
++)
1520 iswhite(i
) = notinname(i
) = begtoken(i
) = intoken(i
) = endtoken(i
) = FALSE
;
1521 for (sp
= white
; *sp
!= '\0'; sp
++) iswhite (*sp
) = TRUE
;
1522 for (sp
= nonam
; *sp
!= '\0'; sp
++) notinname (*sp
) = TRUE
;
1523 notinname('\0') = notinname('\n');
1524 for (sp
= begtk
; *sp
!= '\0'; sp
++) begtoken (*sp
) = TRUE
;
1525 begtoken('\0') = begtoken('\n');
1526 for (sp
= midtk
; *sp
!= '\0'; sp
++) intoken (*sp
) = TRUE
;
1527 intoken('\0') = intoken('\n');
1528 for (sp
= endtk
; *sp
!= '\0'; sp
++) endtoken (*sp
) = TRUE
;
1529 endtoken('\0') = endtoken('\n');
1533 * This routine opens the specified file and calls the function
1534 * which finds the function and type definitions.
1536 node
*last_node
= NULL
;
1539 find_entries (file
, inf
)
1545 node
*old_last_node
;
1547 /* Memory leakage here: the string pointed by curfile is
1548 never released, because curfile is copied into np->file
1549 for each node, to be used in CTAGS mode. The amount of
1550 memory leaked here is the sum of the lengths of the
1552 curfile
= savestr (file
);
1554 /* If user specified a language, use it. */
1556 if (lang
!= NULL
&& lang
->function
!= NULL
)
1559 lang
->function (inf
);
1563 /* Try to guess the language given the file name. */
1564 lang
= get_language_from_filename (file
);
1565 if (lang
!= NULL
&& lang
->function
!= NULL
)
1568 lang
->function (inf
);
1572 /* Look for sharp-bang as the first two characters. */
1573 if (readline_internal (&lb
, inf
) > 0
1575 && lb
.buffer
[0] == '#'
1576 && lb
.buffer
[1] == '!')
1580 /* Set lp to point at the first char after the last slash in the
1581 line or, if no slashes, at the first nonblank. Then set cp to
1582 the first successive blank and terminate the string. */
1583 lp
= etags_strrchr (lb
.buffer
+2, '/');
1587 lp
= skip_spaces (lb
.buffer
+ 2);
1588 cp
= skip_non_spaces (lp
);
1591 if (strlen (lp
) > 0)
1593 lang
= get_language_from_interpreter (lp
);
1594 if (lang
!= NULL
&& lang
->function
!= NULL
)
1597 lang
->function (inf
);
1602 /* We rewind here, even if inf may be a pipe. We fail if the
1603 length of the first line is longer than the pipe block size,
1604 which is unlikely. */
1608 old_last_node
= last_node
;
1609 curlang
= get_language_from_langname ("fortran");
1610 Fortran_functions (inf
);
1612 /* No Fortran entries found. Try C. */
1613 if (old_last_node
== last_node
)
1615 /* We do not tag if rewind fails.
1616 Only the file name will be recorded in the tags file. */
1618 curlang
= get_language_from_langname (cplusplus
? "c++" : "c");
1619 default_C_entries (inf
);
1627 pfnote (name
, is_func
, linestart
, linelen
, lno
, cno
)
1628 char *name
; /* tag name, or NULL if unnamed */
1629 bool is_func
; /* tag is a function */
1630 char *linestart
; /* start of the line where tag is */
1631 int linelen
; /* length of the line where tag is */
1632 int lno
; /* line number */
1633 long cno
; /* character number */
1637 if (CTAGS
&& name
== NULL
)
1640 np
= xnew (1, node
);
1642 /* If ctags mode, change name "main" to M<thisfilename>. */
1643 if (CTAGS
&& !cxref_style
&& streq (name
, "main"))
1645 register char *fp
= etags_strrchr (curfile
, '/');
1646 np
->name
= concat ("M", fp
== NULL
? curfile
: fp
+ 1, "");
1647 fp
= etags_strrchr (np
->name
, '.');
1648 if (fp
!= NULL
&& fp
[1] != '\0' && fp
[2] == '\0')
1653 np
->been_warned
= FALSE
;
1655 np
->is_func
= is_func
;
1657 /* Our char numbers are 0-base, because of C language tradition?
1658 ctags compatibility? old versions compatibility? I don't know.
1659 Anyway, since emacs's are 1-base we expect etags.el to take care
1660 of the difference. If we wanted to have 1-based numbers, we would
1661 uncomment the +1 below. */
1662 np
->cno
= cno
/* + 1 */ ;
1663 np
->left
= np
->right
= NULL
;
1664 if (CTAGS
&& !cxref_style
)
1666 if (strlen (linestart
) < 50)
1667 np
->pat
= concat (linestart
, "$", "");
1669 np
->pat
= savenstr (linestart
, 50);
1672 np
->pat
= savenstr (linestart
, linelen
);
1674 add_node (np
, &head
);
1678 * TAGS format specification
1679 * Idea by Sam Kendall <kendall@mv.mv.com> (1997)
1681 * pfnote should emit the optimized form [unnamed tag] only if:
1682 * 1. name does not contain any of the characters " \t\r\n(),;";
1683 * 2. linestart contains name as either a rightmost, or rightmost but
1684 * one character, substring;
1685 * 3. the character, if any, immediately before name in linestart must
1686 * be one of the characters " \t(),;";
1687 * 4. the character, if any, immediately after name in linestart must
1688 * also be one of the characters " \t(),;".
1690 * The real implementation uses the notinname() macro, which recognises
1691 * characters slightly different form " \t\r\n(),;". See the variable
1694 #define traditional_tag_style TRUE
1696 new_pfnote (name
, namelen
, is_func
, linestart
, linelen
, lno
, cno
)
1697 char *name
; /* tag name, or NULL if unnamed */
1698 int namelen
; /* tag length */
1699 bool is_func
; /* tag is a function */
1700 char *linestart
; /* start of the line where tag is */
1701 int linelen
; /* length of the line where tag is */
1702 int lno
; /* line number */
1703 long cno
; /* character number */
1711 for (cp
= name
; !notinname (*cp
); cp
++)
1713 if (*cp
== '\0') /* rule #1 */
1715 cp
= linestart
+ linelen
- namelen
;
1716 if (notinname (linestart
[linelen
-1]))
1717 cp
-= 1; /* rule #4 */
1718 if (cp
>= linestart
/* rule #2 */
1720 || notinname (cp
[-1])) /* rule #3 */
1721 && strneq (name
, cp
, namelen
)) /* rule #2 */
1722 named
= FALSE
; /* use unnamed tag */
1727 name
= savenstr (name
, namelen
);
1730 pfnote (name
, is_func
, linestart
, linelen
, lno
, cno
);
1735 * recurse on left children, iterate on right children.
1743 register node
*node_right
= np
->right
;
1744 free_tree (np
->left
);
1745 if (np
->name
!= NULL
)
1755 * Adds a node to the tree of nodes. In etags mode, we don't keep
1756 * it sorted; we just keep a linear list. In ctags mode, maintain
1757 * an ordered tree, with no attempt at balancing.
1759 * add_node is the only function allowed to add nodes, so it can
1763 add_node (np
, cur_node_p
)
1764 node
*np
, **cur_node_p
;
1767 register node
*cur_node
= *cur_node_p
;
1769 if (cur_node
== NULL
)
1779 if (last_node
== NULL
)
1780 fatal ("internal error in add_node", (char *)NULL
);
1781 last_node
->right
= np
;
1787 dif
= strcmp (np
->name
, cur_node
->name
);
1790 * If this tag name matches an existing one, then
1791 * do not add the node, but maybe print a warning.
1795 if (streq (np
->file
, cur_node
->file
))
1799 fprintf (stderr
, "Duplicate entry in file %s, line %d: %s\n",
1800 np
->file
, lineno
, np
->name
);
1801 fprintf (stderr
, "Second entry ignored\n");
1804 else if (!cur_node
->been_warned
&& !no_warnings
)
1808 "Duplicate entry in files %s and %s: %s (Warning only)\n",
1809 np
->file
, cur_node
->file
, np
->name
);
1810 cur_node
->been_warned
= TRUE
;
1815 /* Actually add the node */
1816 add_node (np
, dif
< 0 ? &cur_node
->left
: &cur_node
->right
);
1830 /* Output subentries that precede this one */
1831 put_entries (np
->left
);
1833 /* Output this entry */
1837 if (np
->name
!= NULL
)
1838 fprintf (tagf
, "%s\177%s\001%d,%ld\n",
1839 np
->pat
, np
->name
, np
->lno
, np
->cno
);
1841 fprintf (tagf
, "%s\177%d,%ld\n",
1842 np
->pat
, np
->lno
, np
->cno
);
1846 if (np
->name
== NULL
)
1847 error ("internal error: NULL name in ctags mode.", (char *)NULL
);
1852 fprintf (stdout
, "%s %s %d\n",
1853 np
->name
, np
->file
, (np
->lno
+ 63) / 64);
1855 fprintf (stdout
, "%-16s %3d %-16s %s\n",
1856 np
->name
, np
->lno
, np
->file
, np
->pat
);
1860 fprintf (tagf
, "%s\t%s\t", np
->name
, np
->file
);
1864 putc (searchar
, tagf
);
1867 for (sp
= np
->pat
; *sp
; sp
++)
1869 if (*sp
== '\\' || *sp
== searchar
)
1873 putc (searchar
, tagf
);
1876 { /* a typedef; text pattern inadequate */
1877 fprintf (tagf
, "%d", np
->lno
);
1883 /* Output subentries that follow this one */
1884 put_entries (np
->right
);
1887 /* Length of a number's decimal representation. */
1893 while ((num
/= 10) > 0)
1899 * Return total number of characters that put_entries will output for
1900 * the nodes in the subtree of the specified node. Works only if
1901 * we are not ctags, but called only in that case. This count
1902 * is irrelevant with the new tags.el, but is still supplied for
1903 * backward compatibility.
1906 total_size_of_entries (np
)
1914 for (total
= 0; np
!= NULL
; np
= np
->right
)
1916 /* Count left subentries. */
1917 total
+= total_size_of_entries (np
->left
);
1919 /* Count this entry */
1920 total
+= strlen (np
->pat
) + 1;
1921 total
+= number_len ((long) np
->lno
) + 1 + number_len (np
->cno
) + 1;
1922 if (np
->name
!= NULL
)
1923 total
+= 1 + strlen (np
->name
); /* \001name */
1931 #define C_EXT 0x00fff /* C extensions */
1932 #define C_PLAIN 0x00000 /* C */
1933 #define C_PLPL 0x00001 /* C++ */
1934 #define C_STAR 0x00003 /* C* */
1935 #define C_JAVA 0x00005 /* JAVA */
1936 #define C_AUTO 0x01000 /* C, but switch to C++ if `class' is met */
1937 #define YACC 0x10000 /* yacc file */
1940 * The C symbol tables.
1945 st_C_objprot
, st_C_objimpl
, st_C_objend
,
1950 st_C_class
, st_C_template
,
1951 st_C_struct
, st_C_extern
, st_C_enum
, st_C_define
, st_C_typedef
, st_C_typespec
1954 static unsigned int hash
P_((const char *, unsigned int));
1955 static struct C_stab_entry
* in_word_set
P_((const char *, unsigned int));
1956 static enum sym_type C_symtype
P_((char *, int, int));
1958 /* Feed stuff between (but not including) %[ and %] lines to:
1959 gperf -c -k 1,3 -o -p -r -t
1961 struct C_stab_entry { char *name; int c_ext; enum sym_type type; }
1965 while, 0, st_C_ignore
1966 switch, 0, st_C_ignore
1967 return, 0, st_C_ignore
1968 @interface, 0, st_C_objprot
1969 @protocol, 0, st_C_objprot
1970 @implementation,0, st_C_objimpl
1971 @end, 0, st_C_objend
1972 import, C_JAVA, st_C_ignore
1973 package, C_JAVA, st_C_ignore
1974 friend, C_PLPL, st_C_ignore
1975 extends, C_JAVA, st_C_javastruct
1976 implements, C_JAVA, st_C_javastruct
1977 interface, C_JAVA, st_C_struct
1978 class, 0, st_C_class
1979 namespace, C_PLPL, st_C_struct
1980 domain, C_STAR, st_C_struct
1981 union, 0, st_C_struct
1982 struct, 0, st_C_struct
1983 extern, 0, st_C_extern
1985 typedef, 0, st_C_typedef
1986 define, 0, st_C_define
1987 operator, C_PLPL, st_C_operator
1988 template, 0, st_C_template
1989 bool, C_PLPL, st_C_typespec
1990 long, 0, st_C_typespec
1991 short, 0, st_C_typespec
1992 int, 0, st_C_typespec
1993 char, 0, st_C_typespec
1994 float, 0, st_C_typespec
1995 double, 0, st_C_typespec
1996 signed, 0, st_C_typespec
1997 unsigned, 0, st_C_typespec
1998 auto, 0, st_C_typespec
1999 void, 0, st_C_typespec
2000 static, 0, st_C_typespec
2001 const, 0, st_C_typespec
2002 volatile, 0, st_C_typespec
2003 explicit, C_PLPL, st_C_typespec
2004 mutable, C_PLPL, st_C_typespec
2005 typename, C_PLPL, st_C_typespec
2006 # DEFUN used in emacs, the next three used in glibc (SYSCALL only for mach).
2007 DEFUN, 0, st_C_gnumacro
2008 SYSCALL, 0, st_C_gnumacro
2009 ENTRY, 0, st_C_gnumacro
2010 PSEUDO, 0, st_C_gnumacro
2011 # These are defined inside C functions, so currently they are not met.
2012 # EXFUN used in glibc, DEFVAR_* in emacs.
2013 #EXFUN, 0, st_C_gnumacro
2014 #DEFVAR_, 0, st_C_gnumacro
2016 and replace lines between %< and %> with its output,
2017 then make in_word_set static. */
2019 /* C code produced by gperf version 2.7.1 (19981006 egcs) */
2020 /* Command-line: gperf -c -k 1,3 -o -p -r -t */
2021 struct C_stab_entry
{ char *name
; int c_ext
; enum sym_type type
; };
2023 #define TOTAL_KEYWORDS 47
2024 #define MIN_WORD_LENGTH 2
2025 #define MAX_WORD_LENGTH 15
2026 #define MIN_HASH_VALUE 18
2027 #define MAX_HASH_VALUE 138
2028 /* maximum key range = 121, duplicates = 0 */
2035 register const char *str
;
2036 register unsigned int len
;
2038 static unsigned char asso_values
[] =
2040 139, 139, 139, 139, 139, 139, 139, 139, 139, 139,
2041 139, 139, 139, 139, 139, 139, 139, 139, 139, 139,
2042 139, 139, 139, 139, 139, 139, 139, 139, 139, 139,
2043 139, 139, 139, 139, 139, 139, 139, 139, 139, 139,
2044 139, 139, 139, 139, 139, 139, 139, 139, 139, 139,
2045 139, 139, 139, 139, 139, 139, 139, 139, 139, 139,
2046 139, 139, 139, 139, 63, 139, 139, 139, 33, 44,
2047 62, 139, 139, 139, 139, 139, 139, 139, 139, 139,
2048 42, 139, 139, 12, 32, 139, 139, 139, 139, 139,
2049 139, 139, 139, 139, 139, 139, 139, 34, 59, 37,
2050 24, 58, 33, 3, 139, 16, 139, 139, 42, 60,
2051 18, 11, 39, 139, 23, 57, 4, 63, 6, 20,
2052 139, 139, 139, 139, 139, 139, 139, 139, 139, 139,
2053 139, 139, 139, 139, 139, 139, 139, 139, 139, 139,
2054 139, 139, 139, 139, 139, 139, 139, 139, 139, 139,
2055 139, 139, 139, 139, 139, 139, 139, 139, 139, 139,
2056 139, 139, 139, 139, 139, 139, 139, 139, 139, 139,
2057 139, 139, 139, 139, 139, 139, 139, 139, 139, 139,
2058 139, 139, 139, 139, 139, 139, 139, 139, 139, 139,
2059 139, 139, 139, 139, 139, 139, 139, 139, 139, 139,
2060 139, 139, 139, 139, 139, 139, 139, 139, 139, 139,
2061 139, 139, 139, 139, 139, 139, 139, 139, 139, 139,
2062 139, 139, 139, 139, 139, 139, 139, 139, 139, 139,
2063 139, 139, 139, 139, 139, 139, 139, 139, 139, 139,
2064 139, 139, 139, 139, 139, 139, 139, 139, 139, 139,
2065 139, 139, 139, 139, 139, 139
2067 register int hval
= len
;
2073 hval
+= asso_values
[(unsigned char)str
[2]];
2076 hval
+= asso_values
[(unsigned char)str
[0]];
2085 static struct C_stab_entry
*
2086 in_word_set (str
, len
)
2087 register const char *str
;
2088 register unsigned int len
;
2090 static struct C_stab_entry wordlist
[] =
2092 {""}, {""}, {""}, {""}, {""}, {""}, {""}, {""}, {""},
2093 {""}, {""}, {""}, {""}, {""}, {""}, {""}, {""}, {""},
2094 {"if", 0, st_C_ignore
},
2095 {""}, {""}, {""}, {""},
2096 {"int", 0, st_C_typespec
},
2098 {"void", 0, st_C_typespec
},
2100 {"interface", C_JAVA
, st_C_struct
},
2102 {"SYSCALL", 0, st_C_gnumacro
},
2104 {"return", 0, st_C_ignore
},
2105 {""}, {""}, {""}, {""}, {""}, {""}, {""},
2106 {"while", 0, st_C_ignore
},
2107 {"auto", 0, st_C_typespec
},
2108 {""}, {""}, {""}, {""}, {""}, {""},
2109 {"float", 0, st_C_typespec
},
2110 {"typedef", 0, st_C_typedef
},
2111 {"typename", C_PLPL
, st_C_typespec
},
2113 {"friend", C_PLPL
, st_C_ignore
},
2114 {"volatile", 0, st_C_typespec
},
2116 {"for", 0, st_C_ignore
},
2117 {"const", 0, st_C_typespec
},
2118 {"import", C_JAVA
, st_C_ignore
},
2120 {"define", 0, st_C_define
},
2121 {"long", 0, st_C_typespec
},
2122 {"implements", C_JAVA
, st_C_javastruct
},
2123 {"signed", 0, st_C_typespec
},
2125 {"extern", 0, st_C_extern
},
2126 {"extends", C_JAVA
, st_C_javastruct
},
2128 {"mutable", C_PLPL
, st_C_typespec
},
2129 {"template", 0, st_C_template
},
2130 {"short", 0, st_C_typespec
},
2131 {"bool", C_PLPL
, st_C_typespec
},
2132 {"char", 0, st_C_typespec
},
2133 {"class", 0, st_C_class
},
2134 {"operator", C_PLPL
, st_C_operator
},
2136 {"switch", 0, st_C_ignore
},
2138 {"ENTRY", 0, st_C_gnumacro
},
2140 {"package", C_JAVA
, st_C_ignore
},
2141 {"union", 0, st_C_struct
},
2142 {"@end", 0, st_C_objend
},
2143 {"struct", 0, st_C_struct
},
2144 {"namespace", C_PLPL
, st_C_struct
},
2146 {"domain", C_STAR
, st_C_struct
},
2147 {"@interface", 0, st_C_objprot
},
2148 {"PSEUDO", 0, st_C_gnumacro
},
2149 {"double", 0, st_C_typespec
},
2151 {"@protocol", 0, st_C_objprot
},
2153 {"static", 0, st_C_typespec
},
2155 {"DEFUN", 0, st_C_gnumacro
},
2156 {""}, {""}, {""}, {""},
2157 {"explicit", C_PLPL
, st_C_typespec
},
2158 {""}, {""}, {""}, {""}, {""}, {""}, {""}, {""}, {""},
2159 {""}, {""}, {""}, {""}, {""}, {""}, {""}, {""}, {""},
2161 {"enum", 0, st_C_enum
},
2163 {"unsigned", 0, st_C_typespec
},
2164 {""}, {""}, {""}, {""}, {""}, {""}, {""}, {""}, {""},
2165 {"@implementation",0, st_C_objimpl
}
2168 if (len
<= MAX_WORD_LENGTH
&& len
>= MIN_WORD_LENGTH
)
2170 register int key
= hash (str
, len
);
2172 if (key
<= MAX_HASH_VALUE
&& key
>= 0)
2174 register const char *s
= wordlist
[key
].name
;
2176 if (*str
== *s
&& !strncmp (str
+ 1, s
+ 1, len
- 1))
2177 return &wordlist
[key
];
2184 static enum sym_type
2185 C_symtype (str
, len
, c_ext
)
2190 register struct C_stab_entry
*se
= in_word_set (str
, len
);
2192 if (se
== NULL
|| (se
->c_ext
&& !(c_ext
& se
->c_ext
)))
2199 * C functions and variables are recognized using a simple
2200 * finite automaton. fvdef is its state variable.
2204 fvnone
, /* nothing seen */
2205 fdefunkey
, /* Emacs DEFUN keyword seen */
2206 fdefunname
, /* Emacs DEFUN name seen */
2207 foperator
, /* func: operator keyword seen (cplpl) */
2208 fvnameseen
, /* function or variable name seen */
2209 fstartlist
, /* func: just after open parenthesis */
2210 finlist
, /* func: in parameter list */
2211 flistseen
, /* func: after parameter list */
2212 fignore
, /* func: before open brace */
2213 vignore
/* var-like: ignore until ';' */
2216 bool fvextern
; /* func or var: extern keyword seen; */
2219 * typedefs are recognized using a simple finite automaton.
2220 * typdef is its state variable.
2224 tnone
, /* nothing seen */
2225 tkeyseen
, /* typedef keyword seen */
2226 ttypeseen
, /* defined type seen */
2227 tinbody
, /* inside typedef body */
2228 tend
, /* just before typedef tag */
2229 tignore
/* junk after typedef tag */
2233 * struct-like structures (enum, struct and union) are recognized
2234 * using another simple finite automaton. `structdef' is its state
2239 snone
, /* nothing seen yet,
2240 or in struct body if cblev > 0 */
2241 skeyseen
, /* struct-like keyword seen */
2242 stagseen
, /* struct-like tag seen */
2243 sintemplate
, /* inside template (ignore) */
2244 scolonseen
/* colon seen after struct-like tag */
2248 * When objdef is different from onone, objtag is the name of the class.
2250 char *objtag
= "<uninited>";
2253 * Yet another little state machine to deal with preprocessor lines.
2257 dnone
, /* nothing seen */
2258 dsharpseen
, /* '#' seen as first char on line */
2259 ddefineseen
, /* '#' and 'define' seen */
2260 dignorerest
/* ignore rest of line */
2264 * State machine for Objective C protocols and implementations.
2265 * Idea by Tom R.Hageman <tom@basil.icce.rug.nl> (1995)
2269 onone
, /* nothing seen */
2270 oprotocol
, /* @interface or @protocol seen */
2271 oimplementation
, /* @implementations seen */
2272 otagseen
, /* class name seen */
2273 oparenseen
, /* parenthesis before category seen */
2274 ocatseen
, /* category name seen */
2275 oinbody
, /* in @implementation body */
2276 omethodsign
, /* in @implementation body, after +/- */
2277 omethodtag
, /* after method name */
2278 omethodcolon
, /* after method colon */
2279 omethodparm
, /* after method parameter */
2280 oignore
/* wait for @end */
2285 * Use this structure to keep info about the token read, and how it
2286 * should be tagged. Used by the make_C_tag function to build a tag.
2297 } token
; /* latest token read */
2298 linebuffer token_name
; /* its name */
2301 * Variables and functions for dealing with nested structures.
2302 * Idea by Mykola Dzyuba <mdzyuba@yahoo.com> (2001)
2304 static void pushclass_above
P_((int, char *, int));
2305 static void popclass_above
P_((int));
2306 static void write_classname
P_((linebuffer
*, char *qualifier
));
2309 char **cname
; /* nested class names */
2310 int *cblev
; /* nested class curly brace level */
2311 int nl
; /* class nesting level (elements used) */
2312 int size
; /* length of the array */
2313 } cstack
; /* stack for nested declaration tags */
2314 /* Current struct nesting depth (namespace, class, struct, union, enum). */
2315 #define nestlev (cstack.nl)
2316 /* After struct keyword or in struct body, not inside an nested function. */
2317 #define instruct (structdef == snone && nestlev > 0 \
2318 && cblev == cstack.cblev[nestlev-1] + 1)
2321 pushclass_above (cblev
, str
, len
)
2328 popclass_above (cblev
);
2330 if (nl
>= cstack
.size
)
2332 int size
= cstack
.size
*= 2;
2333 xrnew (cstack
.cname
, size
, char *);
2334 xrnew (cstack
.cblev
, size
, int);
2336 assert (nl
== 0 || cstack
.cblev
[nl
-1] < cblev
);
2337 cstack
.cname
[nl
] = (str
== NULL
) ? NULL
: savenstr (str
, len
);
2338 cstack
.cblev
[nl
] = cblev
;
2343 popclass_above (cblev
)
2348 for (nl
= cstack
.nl
- 1;
2349 nl
>= 0 && cstack
.cblev
[nl
] >= cblev
;
2352 if (cstack
.cname
[nl
] != NULL
)
2353 free (cstack
.cname
[nl
]);
2359 write_classname (cn
, qualifier
)
2364 int qlen
= strlen (qualifier
);
2366 if (cstack
.nl
== 0 || cstack
.cname
[0] == NULL
)
2370 cn
->buffer
[0] = '\0';
2374 len
= strlen (cstack
.cname
[0]);
2375 linebuffer_setlen (cn
, len
);
2376 strcpy (cn
->buffer
, cstack
.cname
[0]);
2378 for (i
= 1; i
< cstack
.nl
; i
++)
2383 s
= cstack
.cname
[i
];
2388 linebuffer_setlen (cn
, len
);
2389 strncat (cn
->buffer
, qualifier
, qlen
);
2390 strncat (cn
->buffer
, s
, slen
);
2395 static bool consider_token
P_((char *, int, int, int *, int, int, bool *));
2396 static void make_C_tag
P_((bool));
2400 * checks to see if the current token is at the start of a
2401 * function or variable, or corresponds to a typedef, or
2402 * is a struct/union/enum tag, or #define, or an enum constant.
2404 * *IS_FUNC gets TRUE iff the token is a function or #define macro
2405 * with args. C_EXTP points to which language we are looking at.
2416 consider_token (str
, len
, c
, c_extp
, cblev
, parlev
, is_func_or_var
)
2417 register char *str
; /* IN: token pointer */
2418 register int len
; /* IN: token length */
2419 register int c
; /* IN: first char after the token */
2420 int *c_extp
; /* IN, OUT: C extensions mask */
2421 int cblev
; /* IN: curly brace level */
2422 int parlev
; /* IN: parenthesis level */
2423 bool *is_func_or_var
; /* OUT: function or variable found */
2425 /* When structdef is stagseen, scolonseen, or snone with cblev > 0,
2426 structtype is the type of the preceding struct-like keyword, and
2427 structcblev is the curly brace level where it has been seen. */
2428 static enum sym_type structtype
;
2429 static int structcblev
;
2430 static enum sym_type toktype
;
2433 toktype
= C_symtype (str
, len
, *c_extp
);
2436 * Advance the definedef state machine.
2441 /* We're not on a preprocessor line. */
2442 if (toktype
== st_C_gnumacro
)
2449 if (toktype
== st_C_define
)
2451 definedef
= ddefineseen
;
2455 definedef
= dignorerest
;
2460 * Make a tag for any macro, unless it is a constant
2461 * and constantypedefs is FALSE.
2463 definedef
= dignorerest
;
2464 *is_func_or_var
= (c
== '(');
2465 if (!*is_func_or_var
&& !constantypedefs
)
2472 error ("internal error: definedef value.", (char *)NULL
);
2481 if (toktype
== st_C_typedef
)
2503 if (structdef
== snone
&& fvdef
== fvnone
)
2522 * This structdef business is NOT invoked when we are ctags and the
2523 * file is plain C. This is because a struct tag may have the same
2524 * name as another tag, and this loses with ctags.
2528 case st_C_javastruct
:
2529 if (structdef
== stagseen
)
2530 structdef
= scolonseen
;
2535 && (*c_extp
& C_AUTO
) /* automatic detection of C++ language */
2536 && definedef
== dnone
&& structdef
== snone
2537 && typdef
== tnone
&& fvdef
== fvnone
)
2538 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
2539 if (toktype
== st_C_template
)
2546 && (typdef
== tkeyseen
2547 || (typedefs_or_cplusplus
&& structdef
== snone
)))
2549 structdef
= skeyseen
;
2550 structtype
= toktype
;
2551 structcblev
= cblev
;
2556 if (structdef
== skeyseen
)
2558 structdef
= stagseen
;
2562 if (typdef
!= tnone
)
2565 /* Detect Objective C constructs. */
2575 objdef
= oimplementation
;
2579 case oimplementation
:
2580 /* Save the class tag for functions or variables defined inside. */
2581 objtag
= savenstr (str
, len
);
2585 /* Save the class tag for categories. */
2586 objtag
= savenstr (str
, len
);
2588 *is_func_or_var
= TRUE
;
2592 *is_func_or_var
= TRUE
;
2599 objdef
= omethodtag
;
2600 linebuffer_setlen (&token_name
, len
);
2601 strncpy (token_name
.buffer
, str
, len
);
2602 token_name
.buffer
[len
] = '\0';
2608 objdef
= omethodparm
;
2613 objdef
= omethodtag
;
2614 linebuffer_setlen (&token_name
, token_name
.len
+ len
);
2615 strncat (token_name
.buffer
, str
, len
);
2620 if (toktype
== st_C_objend
)
2622 /* Memory leakage here: the string pointed by objtag is
2623 never released, because many tests would be needed to
2624 avoid breaking on incorrect input code. The amount of
2625 memory leaked here is the sum of the lengths of the
2633 /* A function, variable or enum constant? */
2640 if (fvdef
!= finlist
&& fvdef
!= fignore
&& fvdef
!= vignore
)
2641 fvdef
= fvnone
; /* should be useless */
2649 *is_func_or_var
= TRUE
;
2653 && structdef
== snone
2654 && structtype
== st_C_enum
&& cblev
> structcblev
)
2655 return TRUE
; /* enum constant */
2661 fvdef
= fdefunname
; /* GNU macro */
2662 *is_func_or_var
= TRUE
;
2665 if ((strneq (str
, "asm", 3) && endtoken (str
[3]))
2666 || (strneq (str
, "__asm__", 7) && endtoken (str
[7])))
2671 if ((*c_extp
& C_PLPL
) && strneq (str
+len
-10, "::operator", 10))
2674 *is_func_or_var
= TRUE
;
2677 if (cblev
> 0 && !instruct
)
2679 fvdef
= fvnameseen
; /* function or variable */
2680 *is_func_or_var
= TRUE
;
2691 * C_entries often keeps pointers to tokens or lines which are older than
2692 * the line currently read. By keeping two line buffers, and switching
2693 * them at end of line, it is possible to use those pointers.
2701 #define current_lb_is_new (newndx == curndx)
2702 #define switch_line_buffers() (curndx = 1 - curndx)
2704 #define curlb (lbs[curndx].lb)
2705 #define newlb (lbs[newndx].lb)
2706 #define curlinepos (lbs[curndx].linepos)
2707 #define newlinepos (lbs[newndx].linepos)
2709 #define CNL_SAVE_DEFINEDEF() \
2711 curlinepos = charno; \
2713 linecharno = charno; \
2714 charno += readline (&curlb, inf); \
2715 lp = curlb.buffer; \
2722 CNL_SAVE_DEFINEDEF(); \
2723 if (savetoken.valid) \
2725 token = savetoken; \
2726 savetoken.valid = FALSE; \
2728 definedef = dnone; \
2736 /* This function should never be called when token.valid is FALSE, but
2737 we must protect against invalid input or internal errors. */
2738 if (DEBUG
|| token
.valid
)
2740 if (traditional_tag_style
)
2742 /* This was the original code. Now we call new_pfnote instead,
2743 which uses the new method for naming tags (see new_pfnote). */
2746 if (CTAGS
|| token
.named
)
2747 name
= savestr (token_name
.buffer
);
2748 if (DEBUG
&& !token
.valid
)
2751 name
= concat (name
, "##invalid##", "");
2753 name
= savestr ("##invalid##");
2755 pfnote (name
, isfun
, token
.line
,
2756 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
2759 new_pfnote (token_name
.buffer
, token_name
.len
, isfun
, token
.line
,
2760 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
2761 token
.valid
= FALSE
;
2768 * This routine finds functions, variables, typedefs,
2769 * #define's, enum constants and struct/union/enum definitions in
2770 * C syntax and adds them to the list.
2773 C_entries (c_ext
, inf
)
2774 int c_ext
; /* extension of C */
2775 FILE *inf
; /* input file */
2777 register char c
; /* latest char read; '\0' for end of line */
2778 register char *lp
; /* pointer one beyond the character `c' */
2779 int curndx
, newndx
; /* indices for current and new lb */
2780 register int tokoff
; /* offset in line of start of current token */
2781 register int toklen
; /* length of current token */
2782 char *qualifier
; /* string used to qualify names */
2783 int qlen
; /* length of qualifier */
2784 int cblev
; /* current curly brace level */
2785 int parlev
; /* current parenthesis level */
2786 int typdefcblev
; /* cblev where a typedef struct body begun */
2787 bool incomm
, inquote
, inchar
, quotednl
, midtoken
;
2789 bool yacc_rules
; /* in the rules part of a yacc file */
2790 struct tok savetoken
; /* token saved during preprocessor handling */
2793 initbuffer (&token_name
);
2794 initbuffer (&lbs
[0].lb
);
2795 initbuffer (&lbs
[1].lb
);
2796 if (cstack
.size
== 0)
2798 cstack
.size
= (DEBUG
) ? 1 : 4;
2800 cstack
.cname
= xnew (cstack
.size
, char *);
2801 cstack
.cblev
= xnew (cstack
.size
, int);
2804 tokoff
= toklen
= typdefcblev
= 0; /* keep compiler quiet */
2805 curndx
= newndx
= 0;
2811 fvdef
= fvnone
; fvextern
= FALSE
; typdef
= tnone
;
2812 structdef
= snone
; definedef
= dnone
; objdef
= onone
;
2814 midtoken
= inquote
= inchar
= incomm
= quotednl
= FALSE
;
2815 token
.valid
= savetoken
.valid
= FALSE
;
2818 cplpl
= (c_ext
& C_PLPL
) == C_PLPL
;
2819 cjava
= (c_ext
& C_JAVA
) == C_JAVA
;
2821 { qualifier
= "."; qlen
= 1; }
2823 { qualifier
= "::"; qlen
= 2; }
2831 /* If we're at the end of the line, the next character is a
2832 '\0'; don't skip it, because it's the thing that tells us
2833 to read the next line. */
2854 /* Newlines inside comments do not end macro definitions in
2856 CNL_SAVE_DEFINEDEF ();
2869 /* Newlines inside strings do not end macro definitions
2870 in traditional cpp, even though compilers don't
2871 usually accept them. */
2872 CNL_SAVE_DEFINEDEF ();
2882 /* Hmmm, something went wrong. */
2911 if (fvdef
!= finlist
&& fvdef
!= fignore
&& fvdef
!=vignore
)
2924 else if (/* cplpl && */ *lp
== '/')
2932 if ((c_ext
& YACC
) && *lp
== '%')
2934 /* Entering or exiting rules section in yacc file. */
2936 definedef
= dnone
; fvdef
= fvnone
; fvextern
= FALSE
;
2937 typdef
= tnone
; structdef
= snone
;
2938 midtoken
= inquote
= inchar
= incomm
= quotednl
= FALSE
;
2940 yacc_rules
= !yacc_rules
;
2946 if (definedef
== dnone
)
2949 bool cpptoken
= TRUE
;
2951 /* Look back on this line. If all blanks, or nonblanks
2952 followed by an end of comment, this is a preprocessor
2954 for (cp
= newlb
.buffer
; cp
< lp
-1; cp
++)
2957 if (*cp
== '*' && *(cp
+1) == '/')
2966 definedef
= dsharpseen
;
2967 } /* if (definedef == dnone) */
2973 /* Consider token only if some involved conditions are satisfied. */
2974 if (typdef
!= tignore
2975 && definedef
!= dignorerest
2977 && structdef
!= sintemplate
2978 && (definedef
!= dnone
2979 || structdef
!= scolonseen
))
2985 if (c
== ':' && cplpl
&& *lp
== ':' && begtoken (lp
[1]))
2988 * This handles :: in the middle, but not at the
2989 * beginning of an identifier. Also, space-separated
2990 * :: is not recognised.
2995 goto still_in_token
;
2999 bool funorvar
= FALSE
;
3002 || consider_token (newlb
.buffer
+ tokoff
, toklen
, c
,
3003 &c_ext
, cblev
, parlev
, &funorvar
))
3005 if (fvdef
== foperator
)
3008 lp
= skip_spaces (lp
-1);
3012 && !iswhite (*lp
) && *lp
!= '(')
3015 toklen
+= lp
- oldlp
;
3017 token
.named
= FALSE
;
3018 if ((c_ext
& C_EXT
) /* not pure C */
3019 && nestlev
> 0 && definedef
== dnone
)
3020 /* in struct body */
3022 write_classname (&token_name
, qualifier
);
3023 linebuffer_setlen (&token_name
,
3024 token_name
.len
+qlen
+toklen
);
3025 strcat (token_name
.buffer
, qualifier
);
3026 strncat (token_name
.buffer
,
3027 newlb
.buffer
+ tokoff
, toklen
);
3030 else if (objdef
== ocatseen
)
3031 /* Objective C category */
3033 int len
= strlen (objtag
) + 2 + toklen
;
3034 linebuffer_setlen (&token_name
, len
);
3035 strcpy (token_name
.buffer
, objtag
);
3036 strcat (token_name
.buffer
, "(");
3037 strncat (token_name
.buffer
,
3038 newlb
.buffer
+ tokoff
, toklen
);
3039 strcat (token_name
.buffer
, ")");
3042 else if (objdef
== omethodtag
3043 || objdef
== omethodparm
)
3044 /* Objective C method */
3048 else if (fvdef
== fdefunname
)
3049 /* GNU DEFUN and similar macros */
3051 bool defun
= (newlb
.buffer
[tokoff
] == 'F');
3055 /* Rewrite the tag so that emacs lisp DEFUNs
3056 can be found by their elisp name */
3063 linebuffer_setlen (&token_name
, len
);
3064 strncpy (token_name
.buffer
,
3065 newlb
.buffer
+ off
, len
);
3066 token_name
.buffer
[len
] = '\0';
3069 if (token_name
.buffer
[len
] == '_')
3070 token_name
.buffer
[len
] = '-';
3071 token
.named
= defun
;
3075 linebuffer_setlen (&token_name
, toklen
);
3076 strncpy (token_name
.buffer
,
3077 newlb
.buffer
+ tokoff
, toklen
);
3078 token_name
.buffer
[toklen
] = '\0';
3079 /* Name macros and members. */
3080 token
.named
= (structdef
== stagseen
3081 || typdef
== ttypeseen
3084 && definedef
== dignorerest
)
3086 && definedef
== dnone
3087 && structdef
== snone
3090 token
.lineno
= lineno
;
3091 token
.offset
= tokoff
;
3092 token
.length
= toklen
;
3093 token
.line
= newlb
.buffer
;
3094 token
.linepos
= newlinepos
;
3097 if (definedef
== dnone
3098 && (fvdef
== fvnameseen
3099 || fvdef
== foperator
3100 || structdef
== stagseen
3102 || typdef
== ttypeseen
3103 || objdef
!= onone
))
3105 if (current_lb_is_new
)
3106 switch_line_buffers ();
3108 else if (definedef
!= dnone
3109 || fvdef
== fdefunname
3111 make_C_tag (funorvar
);
3115 } /* if (endtoken (c)) */
3116 else if (intoken (c
))
3122 } /* if (midtoken) */
3123 else if (begtoken (c
))
3134 make_C_tag (TRUE
); /* a function */
3141 if (structdef
== stagseen
&& !cjava
)
3143 popclass_above (cblev
);
3150 if (!yacc_rules
|| lp
== newlb
.buffer
+ 1)
3152 tokoff
= lp
- 1 - newlb
.buffer
;
3157 } /* if (begtoken) */
3158 } /* if must look at token */
3161 /* Detect end of line, colon, comma, semicolon and various braces
3162 after having handled a token.*/
3166 if (yacc_rules
&& token
.offset
== 0 && token
.valid
)
3168 make_C_tag (FALSE
); /* a yacc function */
3171 if (definedef
!= dnone
)
3177 make_C_tag (TRUE
); /* an Objective C class */
3181 objdef
= omethodcolon
;
3182 linebuffer_setlen (&token_name
, token_name
.len
+ 1);
3183 strcat (token_name
.buffer
, ":");
3186 if (structdef
== stagseen
)
3187 structdef
= scolonseen
;
3190 if (definedef
!= dnone
)
3196 make_C_tag (FALSE
); /* a typedef */
3206 if (typdef
== tignore
)
3210 if ((globals
&& cblev
== 0 && (!fvextern
|| declarations
))
3211 || (members
&& instruct
))
3212 make_C_tag (FALSE
); /* a variable */
3215 token
.valid
= FALSE
;
3218 if ((declarations
&& typdef
== tnone
&& !instruct
)
3219 || (members
&& typdef
!= tignore
&& instruct
))
3220 make_C_tag (TRUE
); /* a function declaration */
3226 && structdef
== stagseen
&& (c_ext
& C_PLPL
))
3227 make_C_tag (FALSE
); /* forward declaration */
3229 /* The following instruction invalidates the token.
3230 Probably the token should be invalidated in all other
3231 cases where some state machine is reset prematurely. */
3232 token
.valid
= FALSE
;
3233 } /* switch (fvdef) */
3239 if (structdef
== stagseen
)
3243 if (definedef
!= dnone
)
3249 make_C_tag (TRUE
); /* an Objective C method */
3265 case fvnameseen
: /* a variable */
3266 if ((globals
&& cblev
== 0 && (!fvextern
|| declarations
))
3267 || (members
&& instruct
))
3270 case flistseen
: /* a function */
3271 if ((declarations
&& typdef
== tnone
&& !instruct
)
3272 || (members
&& typdef
!= tignore
&& instruct
))
3274 make_C_tag (TRUE
); /* a function declaration */
3277 else if (!declarations
)
3279 token
.valid
= FALSE
;
3284 if (structdef
== stagseen
)
3288 if (definedef
!= dnone
)
3290 if (structdef
== stagseen
)
3297 make_C_tag (FALSE
); /* a typedef */
3309 if ((members
&& cblev
== 1)
3310 || (globals
&& cblev
== 0
3311 && (!fvextern
|| declarations
)))
3312 make_C_tag (FALSE
); /* a variable */
3321 if (definedef
!= dnone
)
3323 if (objdef
== otagseen
&& parlev
== 0)
3324 objdef
= oparenseen
;
3328 if (typdef
== ttypeseen
3332 /* This handles constructs like:
3333 typedef void OperatorFun (int fun); */
3350 if (definedef
!= dnone
)
3352 if (objdef
== ocatseen
&& parlev
== 1)
3354 make_C_tag (TRUE
); /* an Objective C category */
3368 || typdef
== ttypeseen
))
3371 make_C_tag (FALSE
); /* a typedef */
3374 else if (parlev
< 0) /* can happen due to ill-conceived #if's. */
3378 if (definedef
!= dnone
)
3380 if (typdef
== ttypeseen
)
3382 /* Whenever typdef is set to tinbody (currently only
3383 here), typdefcblev should be set to cblev. */
3385 typdefcblev
= cblev
;
3390 make_C_tag (TRUE
); /* a function */
3399 make_C_tag (TRUE
); /* an Objective C class */
3404 make_C_tag (TRUE
); /* an Objective C method */
3408 /* Neutralize `extern "C" {' grot. */
3409 if (cblev
== 0 && structdef
== snone
&& nestlev
== 0
3416 case skeyseen
: /* unnamed struct */
3417 pushclass_above (cblev
, NULL
, 0);
3420 case stagseen
: /* named struct or enum */
3421 case scolonseen
: /* a class */
3422 pushclass_above (cblev
, token
.line
+token
.offset
, token
.length
);
3424 make_C_tag (FALSE
); /* a struct or enum */
3430 if (definedef
!= dnone
)
3432 if (fvdef
== fstartlist
)
3433 fvdef
= fvnone
; /* avoid tagging `foo' in `foo (*bar()) ()' */
3436 if (definedef
!= dnone
)
3438 if (!noindentypedefs
&& lp
== newlb
.buffer
+ 1)
3440 cblev
= 0; /* reset curly brace level if first column */
3441 parlev
= 0; /* also reset paren level, just in case... */
3445 popclass_above (cblev
);
3447 /* Only if typdef == tinbody is typdefcblev significant. */
3448 if (typdef
== tinbody
&& cblev
<= typdefcblev
)
3450 assert (cblev
== typdefcblev
);
3455 if (definedef
!= dnone
)
3465 if ((members
&& cblev
== 1)
3466 || (globals
&& cblev
== 0 && (!fvextern
|| declarations
)))
3467 make_C_tag (FALSE
); /* a variable */
3474 if (cplpl
&& structdef
== stagseen
)
3476 structdef
= sintemplate
;
3481 if (structdef
== sintemplate
)
3483 structdef
= stagseen
;
3489 if (objdef
== oinbody
&& cblev
== 0)
3491 objdef
= omethodsign
;
3496 case '#': case '~': case '&': case '%': case '/': case '|':
3497 case '^': case '!': case '.': case '?': case ']':
3498 if (definedef
!= dnone
)
3500 /* These surely cannot follow a function tag in C. */
3513 if (objdef
== otagseen
)
3515 make_C_tag (TRUE
); /* an Objective C class */
3518 /* If a macro spans multiple lines don't reset its state. */
3520 CNL_SAVE_DEFINEDEF ();
3526 } /* while not eof */
3528 free (token_name
.buffer
);
3529 free (lbs
[0].lb
.buffer
);
3530 free (lbs
[1].lb
.buffer
);
3534 * Process either a C++ file or a C file depending on the setting
3538 default_C_entries (inf
)
3541 C_entries (cplusplus
? C_PLPL
: C_AUTO
, inf
);
3544 /* Always do plain C. */
3546 plain_C_entries (inf
)
3552 /* Always do C++. */
3554 Cplusplus_entries (inf
)
3557 C_entries (C_PLPL
, inf
);
3560 /* Always do Java. */
3565 C_entries (C_JAVA
, inf
);
3573 C_entries (C_STAR
, inf
);
3576 /* Always do Yacc. */
3581 C_entries (YACC
, inf
);
3585 /* A useful macro. */
3586 #define LOOP_ON_INPUT_LINES(file_pointer, line_buffer, char_pointer) \
3587 for (lineno = charno = 0; /* loop initialization */ \
3588 !feof (file_pointer) /* loop test */ \
3589 && (lineno++, /* instructions at start of loop */ \
3590 linecharno = charno, \
3591 charno += readline (&line_buffer, file_pointer), \
3592 char_pointer = lb.buffer, \
3598 * Read a file, but do no processing. This is used to do regexp
3599 * matching on files that have no language defined.
3602 just_read_file (inf
)
3605 register char *dummy
;
3607 LOOP_ON_INPUT_LINES (inf
, lb
, dummy
)
3612 /* Fortran parsing */
3614 static bool tail
P_((char *));
3615 static void takeprec
P_((void));
3616 static void getit
P_((FILE *));
3622 register int len
= 0;
3624 while (*cp
!= '\0' && lowcase (*cp
) == lowcase (dbp
[len
]))
3626 if (*cp
== '\0' && !intoken (dbp
[len
]))
3637 dbp
= skip_spaces (dbp
);
3641 dbp
= skip_spaces (dbp
);
3642 if (strneq (dbp
, "(*)", 3))
3647 if (!ISDIGIT (*dbp
))
3649 --dbp
; /* force failure */
3654 while (ISDIGIT (*dbp
));
3663 dbp
= skip_spaces (dbp
);
3667 linecharno
= charno
;
3668 charno
+= readline (&lb
, inf
);
3673 dbp
= skip_spaces (dbp
);
3675 if (!ISALPHA (*dbp
) && *dbp
!= '_' && *dbp
!= '$')
3677 for (cp
= dbp
+ 1; *cp
!= '\0' && intoken (*cp
); cp
++)
3679 pfnote (savenstr (dbp
, cp
-dbp
), TRUE
,
3680 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
3685 Fortran_functions (inf
)
3688 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
3691 dbp
++; /* Ratfor escape to fortran */
3692 dbp
= skip_spaces (dbp
);
3695 switch (lowcase (*dbp
))
3698 if (tail ("integer"))
3706 if (tail ("logical"))
3710 if (tail ("complex") || tail ("character"))
3714 if (tail ("double"))
3716 dbp
= skip_spaces (dbp
);
3719 if (tail ("precision"))
3725 dbp
= skip_spaces (dbp
);
3728 switch (lowcase (*dbp
))
3731 if (tail ("function"))
3735 if (tail ("subroutine"))
3743 if (tail ("blockdata") || tail ("block data"))
3745 dbp
= skip_spaces (dbp
);
3746 if (*dbp
== '\0') /* assume un-named */
3747 pfnote (savestr ("blockdata"), TRUE
,
3748 lb
.buffer
, dbp
- lb
.buffer
, lineno
, linecharno
);
3750 getit (inf
); /* look for name */
3760 * Philippe Waroquiers <philippe.waroquiers@eurocontrol.be> (1998)
3763 static void adagetit
P_((FILE *, char *));
3765 /* Once we are positioned after an "interesting" keyword, let's get
3766 the real tag value necessary. */
3768 adagetit (inf
, name_qualifier
)
3770 char *name_qualifier
;
3778 dbp
= skip_spaces (dbp
);
3780 || (dbp
[0] == '-' && dbp
[1] == '-'))
3783 linecharno
= charno
;
3784 charno
+= readline (&lb
, inf
);
3793 /* Skipping body of procedure body or package body or ....
3794 resetting qualifier to body instead of spec. */
3795 name_qualifier
= "/b";
3801 /* Skipping type of task type or protected type ... */
3809 for (cp
= dbp
; *cp
!= '\0' && *cp
!= '"'; cp
++)
3814 dbp
= skip_spaces (dbp
);
3817 && (ISALPHA (*cp
) || ISDIGIT (*cp
) || *cp
== '_' || *cp
== '.'));
3825 name
= concat (dbp
, name_qualifier
, "");
3827 pfnote (name
, TRUE
, lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
3838 bool inquote
= FALSE
;
3840 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
3842 while (*dbp
!= '\0')
3844 /* Skip a string i.e. "abcd". */
3845 if (inquote
|| (*dbp
== '"'))
3847 dbp
= etags_strchr ((inquote
) ? dbp
: dbp
+1, '"');
3852 continue; /* advance char */
3857 break; /* advance line */
3861 /* Skip comments. */
3862 if (dbp
[0] == '-' && dbp
[1] == '-')
3863 break; /* advance line */
3865 /* Skip character enclosed in single quote i.e. 'a'
3866 and skip single quote starting an attribute i.e. 'Image. */
3875 /* Search for beginning of a token. */
3876 if (!begtoken (*dbp
))
3879 continue; /* advance char */
3882 /* We are at the beginning of a token. */
3887 if (!packages_only
&& tail ("function"))
3888 adagetit (inf
, "/f");
3890 break; /* from switch */
3891 continue; /* advance char */
3894 if (!packages_only
&& tail ("procedure"))
3895 adagetit (inf
, "/p");
3896 else if (tail ("package"))
3897 adagetit (inf
, "/s");
3898 else if (tail ("protected")) /* protected type */
3899 adagetit (inf
, "/t");
3901 break; /* from switch */
3902 continue; /* advance char */
3905 if (!packages_only
&& tail ("task"))
3906 adagetit (inf
, "/k");
3907 else if (typedefs
&& !packages_only
&& tail ("type"))
3909 adagetit (inf
, "/t");
3910 while (*dbp
!= '\0')
3914 break; /* from switch */
3915 continue; /* advance char */
3918 /* Look for the end of the token. */
3919 while (!endtoken (*dbp
))
3922 } /* advance char */
3923 } /* advance line */
3928 * Bob Weiner, Motorola Inc., 4/3/94
3929 * Unix and microcontroller assembly tag handling
3930 * look for '^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]'
3938 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
3940 /* If first char is alphabetic or one of [_.$], test for colon
3941 following identifier. */
3942 if (ISALPHA (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
3944 /* Read past label. */
3946 while (ISALNUM (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
3948 if (*cp
== ':' || iswhite (*cp
))
3950 /* Found end of label, so copy it and add it to the table. */
3951 pfnote (savenstr(lb
.buffer
, cp
-lb
.buffer
), TRUE
,
3952 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
3961 * Perl sub names: look for /^sub[ \t\n]+[^ \t\n{]+/
3962 * Perl variable names: /^(my|local).../
3963 * Bart Robinson <lomew@cs.utah.edu> (1995)
3964 * Michael Ernst <mernst@alum.mit.edu> (1997)
3967 Perl_functions (inf
)
3972 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
3976 && *cp
++ == 'b' && iswhite (*cp
++))
3978 cp
= skip_spaces (cp
);
3983 && !iswhite (*cp
) && *cp
!= '{' && *cp
!= '(')
3985 pfnote (savenstr (sp
, cp
-sp
), TRUE
,
3986 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
3989 else if (globals
/* only if tagging global vars is enabled */
3990 && ((cp
= lb
.buffer
,
3999 && (*cp
== '(' || iswhite (*cp
)))
4001 /* After "my" or "local", but before any following paren or space. */
4002 char *varname
= NULL
;
4004 cp
= skip_spaces (cp
);
4005 if (*cp
== '$' || *cp
== '@' || *cp
== '%')
4007 char* varstart
= ++cp
;
4008 while (ISALNUM (*cp
) || *cp
== '_')
4010 varname
= savenstr (varstart
, cp
-varstart
);
4014 /* Should be examining a variable list at this point;
4015 could insist on seeing an open parenthesis. */
4016 while (*cp
!= '\0' && *cp
!= ';' && *cp
!= '=' && *cp
!= ')')
4020 /* Perhaps I should back cp up one character, so the TAGS table
4021 doesn't mention (and so depend upon) the following char. */
4022 pfnote ((CTAGS
) ? savenstr (lb
.buffer
, cp
-lb
.buffer
) : varname
,
4023 FALSE
, lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4031 * Look for /^[\t]*def[ \t\n]+[^ \t\n(:]+/ or /^class[ \t\n]+[^ \t\n(:]+/
4032 * Eric S. Raymond <esr@thyrsus.com> (1997)
4033 * More ideas by seb bacon <seb@jamkit.com> (2002)
4036 Python_functions (inf
)
4041 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4043 cp
= skip_spaces (cp
);
4046 && *cp
++ == 'f' && iswhite (*cp
++))
4048 cp
= skip_spaces (cp
);
4049 while (*cp
!= '\0' && !iswhite (*cp
) && *cp
!= '(' && *cp
!= ':')
4052 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4055 cp
= skip_spaces (lb
.buffer
);
4060 && *cp
++ == 's' && iswhite (*cp
++))
4062 cp
= skip_spaces (cp
);
4063 while (*cp
!= '\0' && !iswhite (*cp
) && *cp
!= '(' && *cp
!= ':')
4066 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4072 /* Idea by Corny de Souza
4073 * Cobol tag functions
4074 * We could look for anything that could be a paragraph name.
4075 * i.e. anything that starts in column 8 is one word and ends in a full stop.
4078 Cobol_paragraphs (inf
)
4081 register char *bp
, *ep
;
4083 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4089 /* If eoln, compiler option or comment ignore whole line. */
4090 if (bp
[-1] != ' ' || !ISALNUM (bp
[0]))
4093 for (ep
= bp
; ISALNUM (*ep
) || *ep
== '-'; ep
++)
4096 pfnote (savenstr (bp
, ep
-bp
), TRUE
,
4097 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
4104 * Idea by Assar Westerlund <assar@sics.se> (2001)
4107 Makefile_targets (inf
)
4112 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4114 if (*bp
== '\t' || *bp
== '#')
4116 while (*bp
!= '\0' && *bp
!= '=' && *bp
!= ':')
4119 pfnote (savenstr (lb
.buffer
, bp
- lb
.buffer
), TRUE
,
4120 lb
.buffer
, bp
- lb
.buffer
+ 1, lineno
, linecharno
);
4125 /* Added by Mosur Mohan, 4/22/88 */
4126 /* Pascal parsing */
4129 * Locates tags for procedures & functions. Doesn't do any type- or
4130 * var-definitions. It does look for the keyword "extern" or
4131 * "forward" immediately following the procedure statement; if found,
4132 * the tag is skipped.
4135 Pascal_functions (inf
)
4138 linebuffer tline
; /* mostly copied from C_entries */
4140 int save_lineno
, save_len
;
4141 char c
, *cp
, *namebuf
;
4143 bool /* each of these flags is TRUE iff: */
4144 incomment
, /* point is inside a comment */
4145 inquote
, /* point is inside '..' string */
4146 get_tagname
, /* point is after PROCEDURE/FUNCTION
4147 keyword, so next item = potential tag */
4148 found_tag
, /* point is after a potential tag */
4149 inparms
, /* point is within parameter-list */
4150 verify_tag
; /* point has passed the parm-list, so the
4151 next token will determine whether this
4152 is a FORWARD/EXTERN to be ignored, or
4153 whether it is a real tag */
4155 save_lcno
= save_lineno
= save_len
= 0; /* keep compiler quiet */
4156 namebuf
= NULL
; /* keep compiler quiet */
4161 initbuffer (&tline
);
4163 incomment
= inquote
= FALSE
;
4164 found_tag
= FALSE
; /* have a proc name; check if extern */
4165 get_tagname
= FALSE
; /* have found "procedure" keyword */
4166 inparms
= FALSE
; /* found '(' after "proc" */
4167 verify_tag
= FALSE
; /* check if "extern" is ahead */
4170 while (!feof (inf
)) /* long main loop to get next char */
4173 if (c
== '\0') /* if end of line */
4176 linecharno
= charno
;
4177 charno
+= readline (&lb
, inf
);
4181 if (!((found_tag
&& verify_tag
)
4183 c
= *dbp
++; /* only if don't need *dbp pointing
4184 to the beginning of the name of
4185 the procedure or function */
4189 if (c
== '}') /* within { } comments */
4191 else if (c
== '*' && *dbp
== ')') /* within (* *) comments */
4208 inquote
= TRUE
; /* found first quote */
4210 case '{': /* found open { comment */
4214 if (*dbp
== '*') /* found open (* comment */
4219 else if (found_tag
) /* found '(' after tag, i.e., parm-list */
4222 case ')': /* end of parms list */
4227 if (found_tag
&& !inparms
) /* end of proc or fn stmt */
4234 if (found_tag
&& verify_tag
&& (*dbp
!= ' '))
4236 /* check if this is an "extern" declaration */
4239 if (lowcase (*dbp
== 'e'))
4241 if (tail ("extern")) /* superfluous, really! */
4247 else if (lowcase (*dbp
) == 'f')
4249 if (tail ("forward")) /* check for forward reference */
4255 if (found_tag
&& verify_tag
) /* not external proc, so make tag */
4259 pfnote (namebuf
, TRUE
,
4260 tline
.buffer
, save_len
, save_lineno
, save_lcno
);
4264 if (get_tagname
) /* grab name of proc or fn */
4269 /* save all values for later tagging */
4270 linebuffer_setlen (&tline
, lb
.len
);
4271 strcpy (tline
.buffer
, lb
.buffer
);
4272 save_lineno
= lineno
;
4273 save_lcno
= linecharno
;
4275 /* grab block name */
4276 for (cp
= dbp
+ 1; *cp
!= '\0' && !endtoken (*cp
); cp
++)
4278 namebuf
= savenstr (dbp
, cp
-dbp
);
4279 dbp
= cp
; /* set dbp to e-o-token */
4280 save_len
= dbp
- lb
.buffer
+ 1;
4281 get_tagname
= FALSE
;
4285 /* and proceed to check for "extern" */
4287 else if (!incomment
&& !inquote
&& !found_tag
)
4289 /* check for proc/fn keywords */
4290 switch (lowcase (c
))
4293 if (tail ("rocedure")) /* c = 'p', dbp has advanced */
4297 if (tail ("unction"))
4302 } /* while not eof */
4304 free (tline
.buffer
);
4309 * Lisp tag functions
4310 * look for (def or (DEF, quote or QUOTE
4313 static int L_isdef
P_((char *));
4314 static int L_isquote
P_((char *));
4315 static void L_getit
P_((void));
4319 register char *strp
;
4321 return ((strp
[1] == 'd' || strp
[1] == 'D')
4322 && (strp
[2] == 'e' || strp
[2] == 'E')
4323 && (strp
[3] == 'f' || strp
[3] == 'F'));
4328 register char *strp
;
4330 return ((*++strp
== 'q' || *strp
== 'Q')
4331 && (*++strp
== 'u' || *strp
== 'U')
4332 && (*++strp
== 'o' || *strp
== 'O')
4333 && (*++strp
== 't' || *strp
== 'T')
4334 && (*++strp
== 'e' || *strp
== 'E')
4335 && iswhite (*++strp
));
4343 if (*dbp
== '\'') /* Skip prefix quote */
4345 else if (*dbp
== '(')
4347 if (L_isquote (dbp
))
4348 dbp
+= 7; /* Skip "(quote " */
4350 dbp
+= 1; /* Skip "(" before name in (defstruct (foo)) */
4351 dbp
= skip_spaces (dbp
);
4354 for (cp
= dbp
/*+1*/;
4355 *cp
!= '\0' && *cp
!= '(' && !iswhite(*cp
) && *cp
!= ')';
4361 pfnote (savenstr (dbp
, cp
-dbp
), TRUE
,
4362 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4366 Lisp_functions (inf
)
4369 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4375 dbp
= skip_non_spaces (dbp
);
4376 dbp
= skip_spaces (dbp
);
4381 /* Check for (foo::defmumble name-defined ... */
4384 while (*dbp
!= '\0' && !iswhite (*dbp
)
4385 && *dbp
!= ':' && *dbp
!= '(' && *dbp
!= ')');
4390 while (*dbp
== ':');
4392 if (L_isdef (dbp
- 1))
4394 dbp
= skip_non_spaces (dbp
);
4395 dbp
= skip_spaces (dbp
);
4406 * Postscript tag functions
4407 * Just look for lines where the first character is '/'
4408 * Also look at "defineps" for PSWrap
4409 * Richard Mlynarik <mly@adoc.xerox.com> (1997)
4410 * Ideas by Masatake Yamato <masata-y@is.aist-nara.ac.jp> (1999)
4413 Postscript_functions (inf
)
4416 register char *bp
, *ep
;
4418 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4423 *ep
!= '\0' && *ep
!= ' ' && *ep
!= '{';
4426 pfnote (savenstr (bp
, ep
-bp
), TRUE
,
4427 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
4429 else if (strneq (bp
, "defineps", 8))
4431 bp
= skip_non_spaces (bp
);
4432 bp
= skip_spaces (bp
);
4440 * Scheme tag functions
4441 * look for (def... xyzzy
4442 * look for (def... (xyzzy
4443 * look for (def ... ((...(xyzzy ....
4444 * look for (set! xyzzy
4448 Scheme_functions (inf
)
4453 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4456 && (bp
[1] == 'D' || bp
[1] == 'd')
4457 && (bp
[2] == 'E' || bp
[2] == 'e')
4458 && (bp
[3] == 'F' || bp
[3] == 'f'))
4460 bp
= skip_non_spaces (bp
);
4461 /* Skip over open parens and white space */
4462 while (iswhite (*bp
) || *bp
== '(')
4467 && (bp
[1] == 'S' || bp
[1] == 's')
4468 && (bp
[2] == 'E' || bp
[2] == 'e')
4469 && (bp
[3] == 'T' || bp
[3] == 't')
4470 && (bp
[4] == '!' || bp
[4] == '!')
4471 && (iswhite (bp
[5])))
4473 bp
= skip_non_spaces (bp
);
4474 bp
= skip_spaces (bp
);
4481 /* Find tags in TeX and LaTeX input files. */
4483 /* TEX_toktab is a table of TeX control sequences that define tags.
4484 Each TEX_tabent records one such control sequence.
4485 CONVERT THIS TO USE THE Stab TYPE!! */
4492 struct TEX_tabent
*TEX_toktab
= NULL
; /* Table with tag tokens */
4494 /* Default set of control sequences to put into TEX_toktab.
4495 The value of environment var TEXTAGS is prepended to this. */
4497 char *TEX_defenv
= "\
4498 :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem\
4499 :part:appendix:entry:index";
4501 static void TEX_mode
P_((FILE *));
4502 static struct TEX_tabent
*TEX_decode_env
P_((char *, char *));
4503 static int TEX_Token
P_((char *));
4505 char TEX_esc
= '\\';
4506 char TEX_opgrp
= '{';
4507 char TEX_clgrp
= '}';
4510 * TeX/LaTeX scanning loop.
4519 /* Select either \ or ! as escape character. */
4522 /* Initialize token table once from environment. */
4524 TEX_toktab
= TEX_decode_env ("TEXTAGS", TEX_defenv
);
4526 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4529 /* Look at each esc in line. */
4530 while ((cp
= etags_strchr (cp
, TEX_esc
)) != NULL
)
4534 linecharno
+= cp
- lasthit
;
4536 i
= TEX_Token (lasthit
);
4539 /* We seem to include the TeX command in the tag name.
4541 for (p = lasthit + TEX_toktab[i].len;
4542 *p != '\0' && *p != TEX_clgrp;
4545 pfnote (/*savenstr (lasthit, p-lasthit)*/ (char *)NULL
, TRUE
,
4546 lb
.buffer
, lb
.len
, lineno
, linecharno
);
4547 break; /* We only tag a line once */
4553 #define TEX_LESC '\\'
4554 #define TEX_SESC '!'
4557 /* Figure out whether TeX's escapechar is '\\' or '!' and set grouping
4558 chars accordingly. */
4565 while ((c
= getc (inf
)) != EOF
)
4567 /* Skip to next line if we hit the TeX comment char. */
4571 else if (c
== TEX_LESC
|| c
== TEX_SESC
)
4587 /* If the input file is compressed, inf is a pipe, and rewind may fail.
4588 No attempt is made to correct the situation. */
4592 /* Read environment and prepend it to the default string.
4593 Build token table. */
4594 static struct TEX_tabent
*
4595 TEX_decode_env (evarname
, defenv
)
4599 register char *env
, *p
;
4601 struct TEX_tabent
*tab
;
4604 /* Append default string to environment. */
4605 env
= getenv (evarname
);
4611 env
= concat (oldenv
, defenv
, "");
4614 /* Allocate a token table */
4615 for (size
= 1, p
= env
; p
;)
4616 if ((p
= etags_strchr (p
, ':')) && *++p
!= '\0')
4618 /* Add 1 to leave room for null terminator. */
4619 tab
= xnew (size
+ 1, struct TEX_tabent
);
4621 /* Unpack environment string into token table. Be careful about */
4622 /* zero-length strings (leading ':', "::" and trailing ':') */
4625 p
= etags_strchr (env
, ':');
4626 if (!p
) /* End of environment string. */
4627 p
= env
+ strlen (env
);
4629 { /* Only non-zero strings. */
4630 tab
[i
].name
= savenstr (env
, p
- env
);
4631 tab
[i
].len
= strlen (tab
[i
].name
);
4638 tab
[i
].name
= NULL
; /* Mark end of table. */
4646 /* If the text at CP matches one of the tag-defining TeX command names,
4647 return the pointer to the first occurrence of that command in TEX_toktab.
4648 Otherwise return -1.
4649 Keep the capital `T' in `token' for dumb truncating compilers
4650 (this distinguishes it from `TEX_toktab' */
4657 for (i
= 0; TEX_toktab
[i
].len
> 0; i
++)
4658 if (strneq (TEX_toktab
[i
].name
, cp
, TEX_toktab
[i
].len
))
4664 /* Texinfo support. Dave Love, Mar. 2000. */
4670 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4676 && *cp
++ == 'e' && iswhite (*cp
++)))
4678 start
= cp
= skip_spaces(cp
);
4679 while (*cp
!= '\0' && *cp
!= ',')
4681 pfnote (savenstr (start
, cp
- start
), TRUE
,
4682 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4689 * Prolog support (rewritten) by Anders Lindgren, Mar. 96
4691 * Assumes that the predicate starts at column 0.
4692 * Only the first clause of a predicate is added.
4694 static int prolog_pred
P_((char *, char *));
4695 static void prolog_skip_comment
P_((linebuffer
*, FILE *));
4696 static int prolog_atom
P_((char *, int));
4699 Prolog_functions (inf
)
4710 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4712 if (cp
[0] == '\0') /* Empty line */
4714 else if (iswhite (cp
[0])) /* Not a predicate */
4716 else if (cp
[0] == '/' && cp
[1] == '*') /* comment. */
4717 prolog_skip_comment (&lb
, inf
);
4718 else if ((len
= prolog_pred (cp
, last
)) > 0)
4720 /* Predicate. Store the function name so that we only
4721 generate a tag for the first clause. */
4723 last
= xnew(len
+ 1, char);
4724 else if (len
+ 1 > allocated
)
4725 xrnew (last
, len
+ 1, char);
4726 allocated
= len
+ 1;
4727 strncpy (last
, cp
, len
);
4735 prolog_skip_comment (plb
, inf
)
4743 for (cp
= plb
->buffer
; *cp
!= '\0'; cp
++)
4744 if (cp
[0] == '*' && cp
[1] == '/')
4747 linecharno
+= readline (plb
, inf
);
4753 * A predicate definition is added if it matches:
4754 * <beginning of line><Prolog Atom><whitespace>(
4756 * It is added to the tags database if it doesn't match the
4757 * name of the previous clause header.
4759 * Return the size of the name of the predicate, or 0 if no header
4763 prolog_pred (s
, last
)
4765 char *last
; /* Name of last clause. */
4770 pos
= prolog_atom (s
, 0);
4775 pos
= skip_spaces (s
+ pos
) - s
;
4777 if ((s
[pos
] == '(') || (s
[pos
] == '.'))
4782 /* Save only the first clause. */
4784 || len
!= (int)strlen (last
)
4785 || !strneq (s
, last
, len
))
4787 pfnote (savenstr (s
, len
), TRUE
, s
, pos
, lineno
, linecharno
);
4795 * Consume a Prolog atom.
4796 * Return the number of bytes consumed, or -1 if there was an error.
4798 * A prolog atom, in this context, could be one of:
4799 * - An alphanumeric sequence, starting with a lower case letter.
4800 * - A quoted arbitrary string. Single quotes can escape themselves.
4801 * Backslash quotes everything.
4804 prolog_atom (s
, pos
)
4812 if (ISLOWER(s
[pos
]) || (s
[pos
] == '_'))
4814 /* The atom is unquoted. */
4816 while (ISALNUM(s
[pos
]) || (s
[pos
] == '_'))
4820 return pos
- origpos
;
4822 else if (s
[pos
] == '\'')
4833 pos
++; /* A double quote */
4835 else if (s
[pos
] == '\0')
4836 /* Multiline quoted atoms are ignored. */
4838 else if (s
[pos
] == '\\')
4840 if (s
[pos
+1] == '\0')
4847 return pos
- origpos
;
4855 * Support for Erlang -- Anders Lindgren, Feb 1996.
4857 * Generates tags for functions, defines, and records.
4859 * Assumes that Erlang functions start at column 0.
4861 static int erlang_func
P_((char *, char *));
4862 static void erlang_attribute
P_((char *));
4863 static int erlang_atom
P_((char *, int));
4866 Erlang_functions (inf
)
4877 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4879 if (cp
[0] == '\0') /* Empty line */
4881 else if (iswhite (cp
[0])) /* Not function nor attribute */
4883 else if (cp
[0] == '%') /* comment */
4885 else if (cp
[0] == '"') /* Sometimes, strings start in column one */
4887 else if (cp
[0] == '-') /* attribute, e.g. "-define" */
4889 erlang_attribute (cp
);
4892 else if ((len
= erlang_func (cp
, last
)) > 0)
4895 * Function. Store the function name so that we only
4896 * generates a tag for the first clause.
4899 last
= xnew (len
+ 1, char);
4900 else if (len
+ 1 > allocated
)
4901 xrnew (last
, len
+ 1, char);
4902 allocated
= len
+ 1;
4903 strncpy (last
, cp
, len
);
4911 * A function definition is added if it matches:
4912 * <beginning of line><Erlang Atom><whitespace>(
4914 * It is added to the tags database if it doesn't match the
4915 * name of the previous clause header.
4917 * Return the size of the name of the function, or 0 if no function
4921 erlang_func (s
, last
)
4923 char *last
; /* Name of last clause. */
4928 pos
= erlang_atom (s
, 0);
4933 pos
= skip_spaces (s
+ pos
) - s
;
4935 /* Save only the first clause. */
4938 || len
!= (int)strlen (last
)
4939 || !strneq (s
, last
, len
)))
4941 pfnote (savenstr (s
, len
), TRUE
, s
, pos
, lineno
, linecharno
);
4950 * Handle attributes. Currently, tags are generated for defines
4953 * They are on the form:
4954 * -define(foo, bar).
4955 * -define(Foo(M, N), M+N).
4956 * -record(graph, {vtab = notable, cyclic = true}).
4959 erlang_attribute (s
)
4965 if (strneq (s
, "-define", 7) || strneq (s
, "-record", 7))
4967 pos
= skip_spaces (s
+ 7) - s
;
4968 if (s
[pos
++] == '(')
4970 pos
= skip_spaces (s
+ pos
) - s
;
4971 len
= erlang_atom (s
, pos
);
4973 pfnote (savenstr (& s
[pos
], len
), TRUE
,
4974 s
, pos
+ len
, lineno
, linecharno
);
4982 * Consume an Erlang atom (or variable).
4983 * Return the number of bytes consumed, or -1 if there was an error.
4986 erlang_atom (s
, pos
)
4994 if (ISALPHA (s
[pos
]) || s
[pos
] == '_')
4996 /* The atom is unquoted. */
4998 while (ISALNUM (s
[pos
]) || s
[pos
] == '_')
5000 return pos
- origpos
;
5002 else if (s
[pos
] == '\'')
5013 else if (s
[pos
] == '\0')
5014 /* Multiline quoted atoms are ignored. */
5016 else if (s
[pos
] == '\\')
5018 if (s
[pos
+1] == '\0')
5025 return pos
- origpos
;
5032 #ifdef ETAGS_REGEXPS
5034 static char *scan_separators
P_((char *));
5035 static void analyse_regex
P_((char *, bool));
5036 static void add_regex
P_((char *, bool, language
*));
5037 static char *substitute
P_((char *, char *, struct re_registers
*));
5039 /* Take a string like "/blah/" and turn it into "blah", making sure
5040 that the first and last characters are the same, and handling
5041 quoted separator characters. Actually, stops on the occurrence of
5042 an unquoted separator. Also turns "\t" into a Tab character.
5043 Returns pointer to terminating separator. Works in place. Null
5044 terminates name string. */
5046 scan_separators (name
)
5050 char *copyto
= name
;
5051 bool quoted
= FALSE
;
5053 for (++name
; *name
!= '\0'; ++name
)
5059 else if (*name
== sep
)
5063 /* Something else is quoted, so preserve the quote. */
5069 else if (*name
== '\\')
5071 else if (*name
== sep
)
5077 /* Terminate copied string. */
5082 /* Look at the argument of --regex or --no-regex and do the right
5083 thing. Same for each line of a regexp file. */
5085 analyse_regex (regex_arg
, ignore_case
)
5089 if (regex_arg
== NULL
)
5091 free_patterns (); /* --no-regex: remove existing regexps */
5095 /* A real --regexp option or a line in a regexp file. */
5096 switch (regex_arg
[0])
5098 /* Comments in regexp file or null arg to --regex. */
5104 /* Read a regex file. This is recursive and may result in a
5105 loop, which will stop when the file descriptors are exhausted. */
5109 linebuffer regexbuf
;
5110 char *regexfile
= regex_arg
+ 1;
5112 /* regexfile is a file containing regexps, one per line. */
5113 regexfp
= fopen (regexfile
, "r");
5114 if (regexfp
== NULL
)
5119 initbuffer (®exbuf
);
5120 while (readline_internal (®exbuf
, regexfp
) > 0)
5121 analyse_regex (regexbuf
.buffer
, ignore_case
);
5122 free (regexbuf
.buffer
);
5127 /* Regexp to be used for a specific language only. */
5131 char *lang_name
= regex_arg
+ 1;
5134 for (cp
= lang_name
; *cp
!= '}'; cp
++)
5137 error ("unterminated language name in regex: %s", regex_arg
);
5141 lang
= get_language_from_langname (lang_name
);
5144 add_regex (cp
+ 1, ignore_case
, lang
);
5148 /* Regexp to be used for any language. */
5150 add_regex (regex_arg
, ignore_case
, NULL
);
5155 /* Turn a name, which is an ed-style (but Emacs syntax) regular
5156 expression, into a real regular expression by compiling it. */
5158 add_regex (regexp_pattern
, ignore_case
, lang
)
5159 char *regexp_pattern
;
5163 static struct re_pattern_buffer zeropattern
;
5166 struct re_pattern_buffer
*patbuf
;
5170 if (regexp_pattern
[strlen(regexp_pattern
)-1] != regexp_pattern
[0])
5172 error ("%s: unterminated regexp", regexp_pattern
);
5175 name
= scan_separators (regexp_pattern
);
5176 if (regexp_pattern
[0] == '\0')
5178 error ("null regexp", (char *)NULL
);
5181 (void) scan_separators (name
);
5183 patbuf
= xnew (1, struct re_pattern_buffer
);
5184 *patbuf
= zeropattern
;
5186 patbuf
->translate
= lc_trans
; /* translation table to fold case */
5188 err
= re_compile_pattern (regexp_pattern
, strlen (regexp_pattern
), patbuf
);
5191 error ("%s while compiling pattern", err
);
5196 p_head
= xnew (1, pattern
);
5197 p_head
->regex
= savestr (regexp_pattern
);
5198 p_head
->p_next
= pp
;
5199 p_head
->language
= lang
;
5200 p_head
->pattern
= patbuf
;
5201 p_head
->name_pattern
= savestr (name
);
5202 p_head
->error_signaled
= FALSE
;
5206 * Do the substitutions indicated by the regular expression and
5210 substitute (in
, out
, regs
)
5212 struct re_registers
*regs
;
5215 int size
, dig
, diglen
;
5218 size
= strlen (out
);
5220 /* Pass 1: figure out how much to allocate by finding all \N strings. */
5221 if (out
[size
- 1] == '\\')
5222 fatal ("pattern error in \"%s\"", out
);
5223 for (t
= etags_strchr (out
, '\\');
5225 t
= etags_strchr (t
+ 2, '\\'))
5229 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
5235 /* Allocate space and do the substitutions. */
5236 result
= xnew (size
+ 1, char);
5238 for (t
= result
; *out
!= '\0'; out
++)
5239 if (*out
== '\\' && ISDIGIT (*++out
))
5242 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
5243 strncpy (t
, in
+ regs
->start
[dig
], diglen
);
5250 assert (t
<= result
+ size
&& t
- result
== (int)strlen (result
));
5255 /* Deallocate all patterns. */
5260 while (p_head
!= NULL
)
5262 pp
= p_head
->p_next
;
5263 free (p_head
->regex
);
5264 free (p_head
->name_pattern
);
5270 #endif /* ETAGS_REGEXPS */
5281 /* Go till you get to white space or a syntactic break */
5283 *cp
!= '\0' && *cp
!= '(' && *cp
!= ')' && !iswhite (*cp
);
5286 pfnote (savenstr (bp
, cp
-bp
), TRUE
,
5287 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5290 /* Initialize a linebuffer for use */
5295 lbp
->size
= (DEBUG
) ? 3 : 200;
5296 lbp
->buffer
= xnew (lbp
->size
, char);
5297 lbp
->buffer
[0] = '\0';
5302 * Read a line of text from `stream' into `lbp', excluding the
5303 * newline or CR-NL, if any. Return the number of characters read from
5304 * `stream', which is the length of the line including the newline.
5306 * On DOS or Windows we do not count the CR character, if any, before the
5307 * NL, in the returned length; this mirrors the behavior of emacs on those
5308 * platforms (for text files, it translates CR-NL to NL as it reads in the
5312 readline_internal (lbp
, stream
)
5314 register FILE *stream
;
5316 char *buffer
= lbp
->buffer
;
5317 register char *p
= lbp
->buffer
;
5318 register char *pend
;
5321 pend
= p
+ lbp
->size
; /* Separate to avoid 386/IX compiler bug. */
5325 register int c
= getc (stream
);
5328 /* We're at the end of linebuffer: expand it. */
5330 xrnew (buffer
, lbp
->size
, char);
5331 p
+= buffer
- lbp
->buffer
;
5332 pend
= buffer
+ lbp
->size
;
5333 lbp
->buffer
= buffer
;
5343 if (p
> buffer
&& p
[-1] == '\r')
5347 /* Assume CRLF->LF translation will be performed by Emacs
5348 when loading this file, so CRs won't appear in the buffer.
5349 It would be cleaner to compensate within Emacs;
5350 however, Emacs does not know how many CRs were deleted
5351 before any given point in the file. */
5366 lbp
->len
= p
- buffer
;
5368 return lbp
->len
+ chars_deleted
;
5372 * Like readline_internal, above, but in addition try to match the
5373 * input line against relevant regular expressions.
5376 readline (lbp
, stream
)
5380 /* Read new line. */
5381 long result
= readline_internal (lbp
, stream
);
5382 #ifdef ETAGS_REGEXPS
5386 /* Match against relevant patterns. */
5388 for (pp
= p_head
; pp
!= NULL
; pp
= pp
->p_next
)
5390 /* Only use generic regexps or those for the current language. */
5391 if (pp
->language
!= NULL
&& pp
->language
!= curlang
)
5394 match
= re_match (pp
->pattern
, lbp
->buffer
, lbp
->len
, 0, &pp
->regs
);
5399 if (!pp
->error_signaled
)
5401 error ("error while matching \"%s\"", pp
->regex
);
5402 pp
->error_signaled
= TRUE
;
5409 /* Match occurred. Construct a tag. */
5410 if (pp
->name_pattern
[0] != '\0')
5412 /* Make a named tag. */
5413 char *name
= substitute (lbp
->buffer
,
5414 pp
->name_pattern
, &pp
->regs
);
5416 pfnote (name
, TRUE
, lbp
->buffer
, match
, lineno
, linecharno
);
5420 /* Make an unnamed tag. */
5421 pfnote ((char *)NULL
, TRUE
,
5422 lbp
->buffer
, match
, lineno
, linecharno
);
5427 #endif /* ETAGS_REGEXPS */
5434 * Return a pointer to a space of size strlen(cp)+1 allocated
5435 * with xnew where the string CP has been copied.
5441 return savenstr (cp
, strlen (cp
));
5445 * Return a pointer to a space of size LEN+1 allocated with xnew where
5446 * the string CP has been copied for at most the first LEN characters.
5455 dp
= xnew (len
+ 1, char);
5456 strncpy (dp
, cp
, len
);
5462 * Return the ptr in sp at which the character c last
5463 * appears; NULL if not found
5465 * Identical to POSIX strrchr, included for portability.
5468 etags_strrchr (sp
, c
)
5469 register const char *sp
;
5472 register const char *r
;
5485 * Return the ptr in sp at which the character c first
5486 * appears; NULL if not found
5488 * Identical to POSIX strchr, included for portability.
5491 etags_strchr (sp
, c
)
5492 register const char *sp
;
5503 /* Skip spaces, return new pointer. */
5508 while (iswhite (*cp
))
5513 /* Skip non spaces, return new pointer. */
5515 skip_non_spaces (cp
)
5518 while (*cp
!= '\0' && !iswhite (*cp
))
5523 /* Print error message and exit. */
5541 suggest_asking_for_help ()
5543 fprintf (stderr
, "\tTry `%s %s' for a complete list of options.\n",
5554 /* Print error message. `s1' is printf control string, `s2' is arg for it. */
5557 const char *s1
, *s2
;
5559 fprintf (stderr
, "%s: ", progname
);
5560 fprintf (stderr
, s1
, s2
);
5561 fprintf (stderr
, "\n");
5564 /* Return a newly-allocated string whose contents
5565 concatenate those of s1, s2, s3. */
5570 int len1
= strlen (s1
), len2
= strlen (s2
), len3
= strlen (s3
);
5571 char *result
= xnew (len1
+ len2
+ len3
+ 1, char);
5573 strcpy (result
, s1
);
5574 strcpy (result
+ len1
, s2
);
5575 strcpy (result
+ len1
+ len2
, s3
);
5576 result
[len1
+ len2
+ len3
] = '\0';
5582 /* Does the same work as the system V getcwd, but does not need to
5583 guess the buffer size in advance. */
5589 char *path
= xnew (bufsize
, char);
5591 while (getcwd (path
, bufsize
) == NULL
)
5593 if (errno
!= ERANGE
)
5597 path
= xnew (bufsize
, char);
5600 canonicalize_filename (path
);
5603 #else /* not HAVE_GETCWD */
5606 char *p
, path
[MAXPATHLEN
+ 1]; /* Fixed size is safe on MSDOS. */
5610 for (p
= path
; *p
!= '\0'; p
++)
5616 return strdup (path
);
5617 #else /* not MSDOS */
5622 pipe
= (FILE *) popen ("pwd 2>/dev/null", "r");
5623 if (pipe
== NULL
|| readline_internal (&path
, pipe
) == 0)
5628 #endif /* not MSDOS */
5629 #endif /* not HAVE_GETCWD */
5632 /* Return a newly allocated string containing the file name of FILE
5633 relative to the absolute directory DIR (which should end with a slash). */
5635 relative_filename (file
, dir
)
5638 char *fp
, *dp
, *afn
, *res
;
5641 /* Find the common root of file and dir (with a trailing slash). */
5642 afn
= absolute_filename (file
, cwd
);
5645 while (*fp
++ == *dp
++)
5647 fp
--, dp
--; /* back to the first differing char */
5649 if (fp
== afn
&& afn
[0] != '/') /* cannot build a relative name */
5652 do /* look at the equal chars until '/' */
5656 /* Build a sequence of "../" strings for the resulting relative file name. */
5658 while ((dp
= etags_strchr (dp
+ 1, '/')) != NULL
)
5660 res
= xnew (3*i
+ strlen (fp
+ 1) + 1, char);
5663 strcat (res
, "../");
5665 /* Add the file name relative to the common root of file and dir. */
5666 strcat (res
, fp
+ 1);
5672 /* Return a newly allocated string containing the absolute file name
5673 of FILE given DIR (which should end with a slash). */
5675 absolute_filename (file
, dir
)
5678 char *slashp
, *cp
, *res
;
5680 if (filename_is_absolute (file
))
5681 res
= savestr (file
);
5683 /* We don't support non-absolute file names with a drive
5684 letter, like `d:NAME' (it's too much hassle). */
5685 else if (file
[1] == ':')
5686 fatal ("%s: relative file names with drive letters not supported", file
);
5689 res
= concat (dir
, file
, "");
5691 /* Delete the "/dirname/.." and "/." substrings. */
5692 slashp
= etags_strchr (res
, '/');
5693 while (slashp
!= NULL
&& slashp
[0] != '\0')
5695 if (slashp
[1] == '.')
5697 if (slashp
[2] == '.'
5698 && (slashp
[3] == '/' || slashp
[3] == '\0'))
5703 while (cp
>= res
&& !filename_is_absolute (cp
));
5705 cp
= slashp
; /* the absolute name begins with "/.." */
5707 /* Under MSDOS and NT we get `d:/NAME' as absolute
5708 file name, so the luser could say `d:/../NAME'.
5709 We silently treat this as `d:/NAME'. */
5710 else if (cp
[0] != '/')
5713 strcpy (cp
, slashp
+ 3);
5717 else if (slashp
[2] == '/' || slashp
[2] == '\0')
5719 strcpy (slashp
, slashp
+ 2);
5724 slashp
= etags_strchr (slashp
+ 1, '/');
5728 return savestr ("/");
5733 /* Return a newly allocated string containing the absolute
5734 file name of dir where FILE resides given DIR (which should
5735 end with a slash). */
5737 absolute_dirname (file
, dir
)
5743 canonicalize_filename (file
);
5744 slashp
= etags_strrchr (file
, '/');
5746 return savestr (dir
);
5749 res
= absolute_filename (file
, dir
);
5755 /* Whether the argument string is an absolute file name. The argument
5756 string must have been canonicalized with canonicalize_filename. */
5758 filename_is_absolute (fn
)
5761 return (fn
[0] == '/'
5763 || (ISALPHA(fn
[0]) && fn
[1] == ':' && fn
[2] == '/')
5768 /* Translate backslashes into slashes. Works in place. */
5770 canonicalize_filename (fn
)
5774 /* Canonicalize drive letter case. */
5775 if (fn
[0] != '\0' && fn
[1] == ':' && ISLOWER (fn
[0]))
5776 fn
[0] = upcase (fn
[0]);
5777 /* Convert backslashes to slashes. */
5778 for (; *fn
!= '\0'; fn
++)
5783 fn
= NULL
; /* shut up the compiler */
5787 /* Set the minimum size of a string contained in a linebuffer. */
5789 linebuffer_setlen (lbp
, toksize
)
5793 while (lbp
->size
<= toksize
)
5796 xrnew (lbp
->buffer
, lbp
->size
, char);
5801 /* Like malloc but get fatal error if memory is exhausted. */
5806 long *result
= (long *) malloc (size
);
5808 fatal ("virtual memory exhausted", (char *)NULL
);
5813 xrealloc (ptr
, size
)
5817 long *result
= (long *) realloc (ptr
, size
);
5819 fatal ("virtual memory exhausted", (char *)NULL
);